HP has been looking for a big, ambitious, risk-it-all project to turn its sinking fortunes around, so it’s perhaps fitting that its new server project is called Moonshot.

Today, HP announced that customers can now buy an HP Moonshot server system. The company claims that Moonshot uses up to 89 percent less energy, 80 percent less space, and will cost 77 percent less than traditional servers.

What does that mean in absolute terms? A single server rack can hold 1,800 Moonshot servers, so you can really cram a lot of the things into a small footprint in a datacenter. Their space consumption is 1/8 that of traditional servers, HP says.

Source: HP

If you prefer your servers with a purse, here’s a stock photo of HP’s Moonshot you might like.

The servers run on Intel’s low-power S1200 processor/system-on-a-chip, which draws just 6 to 10 watts, depending on the chip model. HP claims that, based on initial tests, it will be able to run its entire website, HP.com, using a Moonshot server that draws “the energy equivalency of a dozen 60-watt light bulbs.” HP.com attracts 3 million visits per day, according to the company, so running a site of that scale on just 720 watts is a pretty good feat — it’s the equivalent power draw of just two or three tower PCs.

The first HP Moonshot 1500 server, meant for web hosting, is a 4.3U server enclosure with 45 Intel-based servers inside, plus a network switch and additional components. If you’re not a datacenter ninja, 4.3U means it’s approximately 7 inches high by 19 inches wide and can slide into a server rack, taking up slightly more than 4 standard server slots.

It’s meant for datacenters that need to scale quickly to adapt to changing server loads — such as providers of cloud services.

It will cost $61,875.

Images: Courtesy HP

Electronic Arts has been throwing lots of servers at its connectivity problem for SimCity players, and it now sees light at the end of the suburban sprawl.

Lucy Bradshaw, the general manager of the Maxis label at EA, has become a prolific blogger lately. She said in a post on Sunday that the core problems with getting into the always-connected single-player game are “almost behind us.”

“Our players have been able to connect to their cities in the game for nearly 8 million hours of gameplay time, and we’ve reduced game crashes by 92 percent from day one,” Bradshaw wrote about the city-builder, which is the first major SimCity for the PC in a decade. She said server response time has been improved 40 percent and server capacity has grown 120 percent since launch.

“A combination of optimizing our server architecture and response times, deploying these enhancements on both a series of new and the original servers and issuing a few critical client updates has achieved getting virtually everyone into the game and, once in, having a great time building cities and sharing regions,” she added. “I had hoped to issue an ‘All-Clear’ tonight, but there are still some elements coming together. Tonight and tomorrow we’ll be monitoring each server and gameplay metrics to ensure that the service remains strong and game is playing great. We need a few more days of data before we can assure you that the problem is completely solved and the game is running at 100 percent.”

Bradshaw said that tens of thousands of new players are getting into SimCity every day.

She said the fans have remained confident and that is humbling. It’s taken about six days for EA to show such progress.

“I can’t begin to explain the way a development team feels when something you’re proud of is threatened at launch. Our biggest fear was that people who love this franchise would be scared off by bad reviews about the connectivity issues,” Bradshaw said. “But you put your faith in us. You bought the game with the understanding that we’d quickly fix the server issues. For that support – that incredible commitment from our fans — we are deeply grateful. As the general manager of Maxis, I want you to know that we cherish your faith in us, and the love you’ve shown for this franchise.”

---

GamesBeat 2013 is our fifth annual conference on disruption in the video game market. You'll get 360-degree perspectives from top gaming executives, developers, and analysts on what’s to come in the industry. Our theme this year is “The Battle Royal.” Check out full event details here, and grab your early-bird tickets here!

---

Microsoft and Symantec researchers busted into two data centers in New Jersey and Virginia today to shut down servers associated with a botnet called Bamital.

The companies had an order from the U.S. District Court in Alexandria, Va., according to Reuters, allowing them to enter the data centers. Those at the New Jersey facility seized one of the servers and shut it down. Others in Virginia convinced workers to contact their Netherlands-based parent company and shut down a server there.

The botnet secretly used victim’s computers to steal advertising revenue. The victims didn’t know this was going on until their computers were suddenly unable to search the Internet. Microsoft and Symantec say those people were served with a message that read, “You have reached this website because you computer is very likely to be infected by malware that redirects the results of your search queries. You will receive this notification until you remove the malware from your computer.” It then offered ways to do so.

Bamital is believed to have infected somewhere between 600,000 and 900,000 computers. Microsoft and Symantec are confident they’ve shut down the entire botnet, but notes that only “time will tell,” according to Microsoft’s Digital Crimes Unit’s associate general counsel Richard Boscovich who spoke with Reuters.

This isn’t Microsoft’s first go-around at taking down a botnet, however. In 2011, the company took down the Kelihos botnet, which is small by comparison at on 41,000 infected computers. At the time Microsoft said, however, that the botnet was capable of sending out over 3.8 billion spam emails a day. Microsoft also named suspected perpetrators behind the botnet including a man named Dominique Alexander Piatti.

Later, in 2012, a second and bigger Kelihos botnet was found in the wild. It was subsequently shut down by Russian security firm Kaspersky Lab.

Batman image via kevin dooley/Flickr

Advanced Micro Devices is announcing today that it has designed its standard server motherboard to be more modular and open under the AMD Open 3.0 platform.

That platform, previously code-named Roadrunner,” is a rethinking of the server motherboard to match the standards developed by the Open Compute Project, which is a consortium run by big data center customers such as Facebook and Goldman Sachs as well as server component vendors such as Intel and Arista Networks. AMD hopes the redesign will help it gain market share in server chips, one of the most lucrative computing platforms in the enterprise.

The Open Compute Project was formed to improve the efficiency of servers and to enable data center operators to avoid being locked into solutions provided by component vendors. AMD’s Open 3.0 platform aims to make servers more flexible, efficient, and simple. AMD is targeting markets including high-performance computing, cloud infrastructure, and storage.

“AMD for their 3.0 spec really listened closely to what members were asking for,” said Patrick Moorhead, analyst at Moor Insights & Strategy. “What they came back with was a configurable solution for high performance, general-purpose and storage workloads. AMD really needs a boost in the server space and this could help.”

Many servers are designed with a one-size-fits-most approach. But they can be inefficient when handling unique work loads. AMD is providing tailored solutions with the right combination of power, space and cost, said Suresh Gopalakrishnan, corporate vice president and general manager of AMD’s server business, in an interview with VentureBeat.

“This is going to influence how our customers purchase servers in the future,” he said. “They want their data centers to be as efficient as possible. We are launching the first truly open, modular server. It is exciting for our customers. They are demanding this kind of openness.”

The result will be a lot like putting standard engines or steering wheels into cars. Now application suites will run across different servers from different manufacturers. That gives server administrators more management flexibility.

The AMD solution is currently being evaluated by Fidelity Investments and Goldman Sachs.“This is a realization of the Open Compute Project’s mission of ‘hacking conventional computing infrastructure,’” said Frank Frankovsky, chairman of the Open Compute Foundation and vice president of hardware design and supply chain at Facebook. “What’s really exciting for me here is the way the Open Compute Project inspired AMD and specific consumers to collaboratively bring our ’vanity-free’ design philosophy to a motherboard that suited their exact needs.”

The AMD platform includes the recently announced the recently announced AMD Opteron 6300 Series processors. It can be installed in standard 19-inch racks, and it can fit one, two, or three processor servers. Each server has 12 memory sockets (4 channels with 3 DIMMs each), 6 Serial ATA (SATA) connections per board, a 1 dual channel gigabit Ethernet NIC with integrated management, up to four PCI Express expansion slots, a mezzanine connector for custom module solutions,2 serial ports and 2 USB ports. Specific PCI Express card support is dependent on usage case and chassis height.

The design supports add-on technology from Broadcom and Mellanox. Tyan and Quanta are making the board. More manufacturers are pending.

Google is telling Asian users to expect Google services to run up to 30 percent faster once a trio of new data centers come online.

Google’s head of products in India, Lalitesh Katragadda, told India’s Economic Times that it expects data centers in Singapore and Taiwan to be operational in 2013. A third data center in Hong Kong is also under construction.

As a result, Google services users in most parts of Asian, including the Indian subcontinent, can expect faster, more responsive performance from their Google-made web apps.

While India is one of Google’s largest markets, Katragadda said the climate wasn’t suitable for data centers (odd, considering Google competitor Facebook has had such marked success in hotter weather and higher temperatures). Still, the company hopes to boost slow speeds for services like YouTube and Gmail with relatively nearby facilities.

“Internet connectivity speed in India is not very high. These data centers will be crucial to this market due to its proximity,” said Katragadda.

“More new Internet users are coming online everyday here in Asia than anywhere else in the world,” said Google of its new facility in Changhua County, Taiwan. The entire project, which it expects to complete in the second half of 2013, will cost around $300 million by the time it’s finished.

“This data center will be the first in our fleet to save energy through a nighttime cooling and thermal energy storage system,” reads the project’s page. “And, like our other facilities in Asia, this will be one of the most efficient and environmentally friendly data centers in the region, built to the same high standard we use around the world.”

The Hong Kong data center is still in formative stages, just more than a year after Google bought the 2.7 hectares of land in the Tseung Kwan O Industrial Estate in Kowloon on which it’s constructing the new facility. The Singapore facility will be the first of the three to start operations, flipping its switches in the next couple months.

Interestingly, Google is launching community grants in each area where the new data centers will be opened.

Generally, Google claims its data centers use 50 percent less energy than those of its competitors, which for various products include Facebook, Amazon, Yahoo, Microsoft, and dozens of other companies.

top image via Google

---

Big Data and Predictive/Real-time Analytics startups: Are you looking to jumpstart development & accelerate market traction? Sign up for the SAP Startup Focus program to receive technology, support, resources and community to help you develop new applications on SAP HANA, a cutting edge database platform. Get started here, and enter promo code “VB2013″ on the form.

---

Nobody likes downtime, but while your servers might be out of commission, your revenue could be dropping, too. In fact, businesses lost about $70 million in the past five years just by having intermittent downtime.

There’s a steep cost for letting servers go dark every once in awhile. Even if the blip is just that, a quick moment when you can’t access anything, you could be losing out on sales and losing data. According to the infographic created by MegaPath below, if you are a small to medium-sized business, and if your systems run at even 99.5 percent efficiency, you could be losing up to $12,500 an hour.

And in that time offline, one in every two small businesses experience some kind of data loss.

Downtime is most often caused by power station problems, followed by hardware failure, network failure, floods, and human error. These failures could be the cause of internal issues, or hacks.

The goal efficiency rate is 99.999 percent up-time. Where are your servers?

Created by MegaPath

Orange cones image via Shutterstock

That’s a bit of hyperbole, of course, because even though the world’s largest search engine indexes perhaps 50 billion web pages, the Internet itself is more than just Google.

But Google’s millions of servers are certainly “one of the most powerful server networks in the known Universe,” as Google says, and almost certainly the most powerful server network on the planet. And today the company is giving the universe a sneak peek inside with a cool online picture gallery.

That gallery, surprisingly, is short on information, saying little about the actual technical details, such as the probably 2 million servers Google runs, including almost half a million at its Georgia data center. Or how many billions of miles of Ethernet cable the company uses.

But it does have a very cool view of the places your computer has already been via Google. Apparently, while I’m sitting here with my butt in my chair, my MacBook Air may be virtually visiting Hamina, Finland, St. Ghislain, Belgium, and assorted U.S. destinations like Iowa and Oklahoma.

And Google does give us some amazing pictures of the machines that connect our lives.

Check them out:

Image credits: Google

Today, Facebook is showing off an internal tool it built to help it quickly find and fix caching issues.

Engineers can use Claspin*, built by Facebooker Sean Lynch, to scan and suss out performance problems from TCP retransmits to timeouts using a heatmap that makes troubleshooting much simpler.

Lynch noted in a blog post today that when he first came to the company, Facebook was working with two caching systems, Memcache and TAO, with thousands of charts and an array of dashboards within the company’s operations data store.

“This worked well at first, but as Facebook grew both in size and complexity, it became more and more difficult to figure out which piece was broken when something went wrong,” said Lynch.

So the engineer thought of heatmaps as a better way to visualize the data, with each square in the heatmap representing a host and with racks grouped together. The heatmaps shows green squares for fully functioning hosts and red squares for hosts with problems. Yellow squares show a metric heading into a less-than-optimal range.

Also, said Lynch, the visual display relates to the racks’ physical layout. “The rack names naturally sort by datacenter, then cluster, then row, so problems common at any of these levels are readily apparent.”

With Claspin up and running, Lynch continued, “On a 30-inch screen, we could easily fit 10,000 hosts at the same time, with 30 or more stats contributing to their color, updated in real time.”

Here’s what Claspin looks like and more about how it works:

*Fascinating side note: Claspin is named after CLSPN, a protein-coding gene in the human genome that acts as a sensor to monitor the integrity of DNA replication forks, checking for damage and replication.

---

Big Data and Predictive/Real-time Analytics startups: Are you looking to jumpstart development & accelerate market traction? Sign up for the SAP Startup Focus program to receive technology, support, resources and community to help you develop new applications on SAP HANA, a cutting edge database platform. Get started here, and enter promo code “VB2013″ on the form.

---

Sept. 9 - 10, 2013
San Francisco, CA

Early Bird Tickets on Sale

We are excited to announce the first breakout panel at CloudBeat 2012, “Visibility: Cloud Management and Control in a Complex World,” featuring speakers from 6 Fusion, Cloudability and Appescute. 

Even when you buy a car with a warranty, you still have to get it serviced. Your cloud infrastructure is no different.

“It’s tempting to think of [the cloud] as something that’s already taken care of,” said Joshua Greenbaum, principal at Enterprise Applications Consulting, a Bay Area-based research firm. “Even if you’re a startup, you may still have to do a tremendous amount of sophisticated monitoring.”

Customers are notoriously fickle, and for digital natives, failure is not an option. From a small garage startup to a large enterprise, IT is under intense pressure to spot networking problems in seconds — before they disrupt customer service. That’s especially true for startups and small to medium-sized businesses (SMBs).

“For startups without much of an existing reputation, it’s even more critical to have reliable uptime for your website and services,” said Donnie Berkholz, an analyst at Redmonk, a developer-focused research firm. Berkholz stressed that the dangers for SMBs and young tech companies are very real, even if they’re not hosting a mission-critical application or providing a mission-critical service.

Cloud monitoring is a necessity, but the market is flooded with options. So VentureBeat asked a host of experts to cut to the chase: What are the best cloud-monitoring services out there?

Well, options for cloud monitoring tools will vary depending on the size of your company.

Small to medium sized businesses

Pingdom, a free web monitoring tool, is the most lightweight cloud monitoring solution on the market — it’s a simple ‘heads-up’ alert system that is popular with startups and SMBs. (Tip: Next time Twitter is down, you can track its performance on Pingdom.)

A simple monitoring service is not a cloud-management tool, but it’s a start. For more full-fledged options, you have a number of choices.

Rackspace recently announced cloud monitoring tools designed to let businesses monitor their IT infrastructures, regardless of whether they’re hosted in the cloud or on-premise. For Greenbaum, this is the most viable choice for startups and SMBs that are already using Rackspace for their cloud hosting.

“Our competitive advantage is how well it integrates with the rest of Rackspace,” said Pat Mathews, senior vice president of corporate development at Rackspace. “Beyond that it’s scalable and built for the new workloads.” Austin, Texas-based Rackspace offers a portfolio of cloud products and picked up Cloudkick in 2010, which forms the basis of the new monitoring tool. It works by sending customers an alert if things are going wrong with equipment, irrespective of where it’s hosted.

Of course, Rackspace isn’t the only choice.

Ray Wang, CEO at Constellation Research Group and an expert on all things cloud-related, says your decision depends not only on the size of your business, but also the vertical you’re in.

For e-commerce companies, Wang suggests AlertSite, an applications performance management solution with 100,000 users, acquired last year by Smartbear software.

“What you want is rich tools, easy analytics, and the ability to transcend cloud and on-premises,” he said.

Copper Egg, he told me, is another good solution for those that have opted for hybrid cloud environments. Copper Egg also won the approval of Berkholz.

Berkholz’s other top picks are Boundary, which offers speedier monitoring for hybrid cloud environments, and Zenoss, an open source option.

For larger enterprises

IT departments understand that legacy IT monitoring solutions are riddled with problems: They were never developed to deal with hybrid cloud environments. They are comprised of piecemeal technologies, originally designed to gain insight into whether services and applications are living up to their service-level agreements (SLAs) and flag IT when there are service interruptions.

Trying to configure a legacy solution to extend monitoring beyond the traditional data center presents a new set of cost, learning, and reporting challenges for CIO’s, as eWeek reports.

And when it comes to the big-boy vendors in the space, competition is intense. The experts throw up a smorgasbord of names, with BMC, Nimsoft by CA, and ScienceLogic popping up most often.

Amazon Web Services (AWS) is by far the largest provider of cloud-based infrastructure services. It competes with Microsoft Azure, Joyent, and Rackspace. Amazon has its own cloud monitoring tool called CloudWatch, which it offers to premium customers. If you’re a big user of AWS, this is one option to investigate.

This is a decision best left in the hands of IT. However, here is a series of questions big companies can ask before making a decision about a cloud monitoring provider:

1. Is the architecture scalable? Can it deliver monitoring that meets the expectations of the organization’s growth?
2. Does it deliver on-demand monitoring? Can it respond in a matter of seconds?
3. Does it meet your budgetary requirements? Are there flexible purchasing options?
4. Are there customizable dashboards that grant you visibility?
5. How speedy is the deployment?

It’s still not yet clear whether large-scale enterprises will stop dragging their feet and make the jump to a full cloud-based infrastructure. One way or another, they should be on the look-out for a monitoring tool that can keep up with the pace. It’s all about agility, flexibility, and speed.

As Gary Read, CEO of Boundary and formerly an executive at Nimsoft, put it, “Most IT management tools were designed in a different era when infrastructure was static. It’s no longer acceptable to take a look at your applications every five minutes.” That would be a massive fail whale.

Where to learn more

If you’re stuck, Christian Dawson, chairman of the board of directors at the Internet Infrastructure Coalition, recommends sites like WebHostingTalk.com, where you can ask peers and experts for their thoughts about various companies’ capabilities. ”Nobody is going to have perfect feedback — the market is still too young, fragmented, and misunderstood, but you’ll be able to tell pretty quickly who is reliable and responsive,” he said.

If you’re new to all this, before you go shopping, the first decision you’ll need to make is a structural one. Will you build your monitoring in-house using open source software, or will you outsource to an existing company? The majority of the experts recommend the latter.

Finally, ask the sales guy the tough questions around things like actionable alerts. In a recent interview with Rackspace to discuss the launch of its cloud monitoring tool set, Pat Matthews, senior vice president of corporate development, admitted that this is a notoriously tricky problem for vendors. Some problems need to be rectified by a developer, others will fix themselves. So which alerts are truly actionable? “We are starting down that path,” he said. “It’s something we will continually strive to improve.”

The good news is that there isn’t a world of difference between the various solutions on the market. Of course, there are the legacy, best of breed solutions that EMC, HP and the like offer. But if you’re not running a large enterprise, you have plenty of cost-effective alternatives.

“With rare exceptions, we’re talking about very similar kinds of products,” said Greenbaum. He used the analogy of a heart monitoring system. “You don’t care about the brand or the packaging,” he said. “You just need something that works.”

Security camera image via Shutterstock

SAN FRANCISCO — When Intel warned about weak third-quarter sales last week, it blamed it on slow consumer and enterprise PC sales. But it said that data center computers known as servers were meeting expectations. The reason is that the explosion of tablets, smartphones, and other mobile devices helps drive demand for servers, said Diane Bryant, the head of data center computing at Intel.

Speaking at the Intel Developer Forum in San Francisco today, Bryant (pictured above) said that for every 120 tablets sold, a server has to exist to handle all of the web-based traffic that their users create. She said that for every 20 digital signs in a shopping mall and other areas, like the flatscreens in your local coffee shop, you need one server. And for every 20 surveillance cameras deployed, you need a server to sort through all of the video data.

Servers are pretty universal computers that are used in connected systems, small businesses, communications infrastructure, enterprises, cloud services, technical computing, and mission-critical computing centers like stock markets. As the era of “big data” arrives, companies must deploy big data centers to process it.

Intel is sampling its next-generation E7 and E5 server chips for delivery next year. The low-power server chips are based on Intel’s Ivy Bridge chip designs, which combine graphics and computing on a single chip.

Low-power Xeon E3 series chips arrive in 2013 and are based on the 22-nanometer Haswell core now in development. But a new breed of lower-power Atom-based servers, known as microservers, are also arriving. These microservers have chips with a third of the usual power consumption and they use Intel’s Atom processors. Next year, Intel will launch the 22-nanometer Avoton version of the Atom chip for microservers.

BMW said it will use a new Open Data Center Alliance data center, fueled by geothermal and hydroelectric energy, at a location in Iceland.

On the security front, Bryant said that Intel is embedding Deep Defender security technology into its Xeon server chips. Intel has also launched trusted execution technology for its ecosystem. Intel entered the server chip market in 1997 and has improved performance 10,000 times.

Big data was a $4 billion market last year, Bryant said, and it is growing at 30 percent a year. The data comes from everything, like social networks or radio frequency identification sensors.

Ariel Kelman, head of worldwide marketing at Amazon’s web services division, which hosts web infrastructure for businesses, said that it is working closely with Intel to transform supercomputing from a tool for the elite to something the masses can tap into. Amazon’s Cluster Compute Eight Extra Large Instances are ideal for big data server solutions. They start at $2.40 per hour. You can configure your own top 500 supercomputer with 290 instances at a cost of $73 per hour.

Yelp uses AWS to analyze all of the data from clicks on its online user reviews. NASA uses AWS to calculate telemetry data for the Mars Curiosity rover.

Six Flags uses Intel servers in its amusement parks to monitor crowd traffic via surveillance cameras in its theme parks. It reallocates its staff to the high-traffic areas based on the data from the cameras.

Back in the spring, it was no surprise that AMD would still be selling SeaMicro’s microservers with Intel chips. But today, AMD’s SeaMicro division launched its brand new SeaMicro SM15000 server, which runs with Intel’s E3-1260L Sandy Bridge-based processor. In November, AMD will start shipping new machines with a new generation of AMD Opteron processors with AMD’s Piledriver core as well as Intel’s Xeon E3-1265Lv2 Ivy Bridge processor. In other words, AMD is accommodating SeaMicro’s existing customers by shipping whatever processor that want, even if that gives some business to Intel.

SeaMicro’s small and power-efficient computers enable enterprises to cram more computing power into a given amount of space and use a lot less electrical power. That cuts electricity bills, the largest cost of operating a data center. Using Intel Atom processors, SeaMicro had become the fastest-growing system company in Silicon Valley history before AMD bought it. SeaMicro’s customers include France Telecom, Skype, Rogers Wireless, Mozilla, eHarmony, and China Netcom BB. Hundreds of millions of internet users traverse SeaMicro’s hardware daily.

As we’ve described in earlier stories, Intel has been improving its server microprocessors by making them more power efficient. But the microprocessor only accounts for a third of the power consumption in a server. SeaMicro’s innovation lies in how it attacks the remaining two-thirds of the power consumption problem. It does so by combining a lot of the extraneous chips into a single, more-efficient custom chip. With the low-power cores, SeaMicro can now fit thousands of cores and also 5 petabytes of storageinto a single microserver.

“The rise of virtualization, cloud computing, and big data require a new generation of compute in which networking and storage are equal partners in the solution. This does not fit the mold of traditional servers,” said Andrew Feldman, general manager and corporate vice president of the Data Center Server Solutions group, AMD. “We are at the beginning of a new wave of computing that requires data centers to become pools of computing and storage resources with the flexibility to expand in both dimensions.  The SM15000 system removes the constraints of traditional servers and allows data centers to expand compute, networking and storage independently. By supporting the newest generation of processors, the SM15000 server will continue our tradition of being the highest-density, and most power efficient micro server in the market.”

The new SeaMicro SM15000 server can fit 512 cores in a ten-rack system. Each system has 64 slots for server cards. Diane Bryant, head of Intel’s data center server chip business, said about the AMD use of Intel server chips, “We love all customers. We are happy to sell Intel Xeon chips to anyone.”

CloudBeat 2012 is assembling the biggest names in the cloud’s evolving story to learn about real cases of revolutionary cloud adoption. Unlike other cloud events, customers — the users of cloud technologies — will be front and center. Their discussions with vendors and other experts will give you rare insights into what really works, who’s buying what, and where the industry is going. Register now and save 25 percent! The early-bird discount ends September 14.

President Obama’s Ask Me Anything two days ago was a massive success, as Reddit highlighted earlier today in a blog post. Not only did the page get 2.99 million views on the day of the event, it has received another 2.3 million page views already as of this morning.

Perhaps the most surprising stat was the number of servers Reddit added to handle the onslaught of traffic. In preparation, Reddit added 30 dedicated servers “just for the comment thread.” When that turned out to not be enough, developers added another 30 dedicated servers.

Besides telling you how popular and/or controversial President Obama is, that also tells you that Reddit can scale servers very, very quickly.

More interesting stats from Reddit:

• Peak transfer: 48 MBs per second
• Peak traffic: over 100,000 page views per minute
• Peak concurrent visitors: 198,000
• Peak share of Reddit traffic to the AMA page: 30 percent (average for a very hot topic is 2-5 percent)

In fact, Obama’s AMA is the “only link in recorded history to have surpassed the front page.”

Reddit’s popularity seems to be growing with no end in sight, with between two and eight million unique visitors every month and six to eight million pageviews a day. Obama answering questions on the site will certainly increase that trend and expose the site to a wider, more mainstream audience.

Obama’s most interesting comment? The recipe for White House Beer will be available shortly.

And from the president’s own tongue: it’s tasty.

photo credit: Chuckumentary via photo pin cc

UPDATE: Facebook may end up open-sourcing the new cold-storage servers, too. See below.

Facebook’s Sub-Zero data center is a wee little facility compared to its next-door neighbor, the huge Prineville, Ore., data center that houses the social network’s open-source servers.

But Sub-Zero isn’t for serving up Likes at lightning speed; it’s for long-term, low-power backups that will keep Facebook data on ice indefinitely. This is referred to as cold storage — hanging on to data that you can’t delete but don’t need to get at very often.

Facebook engineers are still working on the designs for the servers and the custom-tweaked software that will live in Sub-Zero; these won’t be out until sometime next year. And Facebook VP Tom Furlong revealed to Wired today that Facebook won’t be accessing the devices too frequently; they’ll be specifically for deep storage.

While other companies still use tape for such backups, Facebook will be using these all-digital new devices. Currently, the company keeps one backup for use in the event of server issues and one backup for emergencies.

Sub-Zero is just 62,000 square feet, one fifth the size of the main Prineville facility. This will be Facebook’s third building in Prineville.

Given the Prineville center’s focus on Open Compute and open-source hardware, we wondered whether Facebook plans to open-source its new deep-storage servers. A company spokesperson tells us that open-sourcing the new devices certainly would fit with Facebook’s current modus operandi; however, as the devices are still in the design stage, it’s too early to release any details just yet.

Amazon, Apple, Dell, Google, Microsoft, and Yahoo also operate data facilities in the same general area of Oregon; it’s nice and cool, and they all get incentivized to bring short-term and long-term jobs to the area.

Image of Facebook’s Prineville data center courtesy of Jolie O’Dell, Flickr

---

Big Data and Predictive/Real-time Analytics startups: Are you looking to jumpstart development & accelerate market traction? Sign up for the SAP Startup Focus program to receive technology, support, resources and community to help you develop new applications on SAP HANA, a cutting edge database platform. Get started here, and enter promo code “VB2013″ on the form.

---

I talked to Sam Parnell, Bleacher Report’s vice president of engineering, about how he’s preparing for the onslaught.

“We use Ruby on Rails,” Parnell told VentureBeat. “It has a bit of a bad rap as not able to scale, but we’re proof that it can.”

Bleacher Report starts by heavily caching everything that can be cached: images, database content, queries. Caching reduces database fetches and asset retrievals, so the site actually speeds up when the company hits high traffic periods. That’s due to the fact that traffic peaks are almost always associated with single events that are topical and important, reducing the number of assets that need to be cached.

Traffic peaks like that are a good thing for Bleacher Report. Because of its crowdsourced content model, usually the site’s traffic is to very diverse topics.

“What makes us different is how we create content. Most of our competitors have 25-35 journalists and focus on national coverage or popular teams like the New York Yankees,” Parnell said. “We have 2,000 featured contributors, which allows us to generate more sports coverage than anybody else … up to 1,000 stories a day.”

Those thousand stories generate immense traffic, upwards of 90 million monthly visits, and the immense traffic generates massive server load.

To monitor that load, Bleacher Report uses New Relic, an application performance management tool. A somewhat unique aspect of the site’s development model is that engineers continue working on features for their entire lifecycle, from development to production. Every engineer has New Relic open almost full-time, monitoring their parts of the overall site, constantly tuning and optimizing.

“We also spend a lot of time optimizing queries,” Parnell says. “It’s a never-ending task.”

Because Bleacher Report runs on Amazon Web Services, adding additional capacity can be very quick. Parnell can add 50 percent more web serving capacity in 30 minutes.

But just because it’s cloud does not mean that everything is super-simple: Database resources take much more time and configuration to increase. So another thing Parnell does is proactively pre-provision excess database resources.

“Actually, this week we’re scaling up our database layer already,” he told VentureBeat.

The previous record-breaking traffic for Bleacher Report, which peaked at 80,000 requests/minute, was during the Floyd Merryweather-Miguel Cotto boxing match in May. Parnell is expecting three times more for the Olympics. If Bleacher Report makes it, perhaps they’ll start to close with ESPN and Yahoo Sports’ traffic numbers.

Bleacher Report was founded in 2006, and has received about $35 million in funding from Oak Investment Partners, Crosslink Capital, and Hillsven Capital.

Image credit: Pete Saloutos/ShutterStock

Sept. 9 - 10, 2013
San Francisco, CA

Early Bird Tickets on Sale

With so much talk about “big data” lately and data-focused companies like 10gen and Delphix recently grabbing large funding rounds, it’s a topic that won’t be going away any time soon. But what exactly is it and is there an easier way to comprehend just how big it all is?

Typically, big data describes working with monster-sized data sets that are hard to manage using standard database tools. Thus, we have all kinds of companies scrambling to provide new software and tools to manage, store, and analyze those data sets. Other big data startups that have seen investments include Cloudera, Palantir, Datahero, and Precog. On top of that, big dogs like Google, Yahoo, and HP are providing their own solutions too.

HP-owned Autonomy has taken a stab at trying to explain “big data” with a new infographic. Autonomy’s private cloud surpassed 50 petabytes in April and just last week, the company released new tools focused on big data, so clearly the company has big data on its mind.

Check out the infographic below:

Top photo credit: VLADGRIN/Shutterstock

What do HP, Salesforce, AMD, VMWare, and Alibaba all have in common?

They’re all partners in the Facebook-led Open Compute Project, a group that aims to revolutionize computer hardware through the power of open-source collaboration.

Facebook’s Open Compute Project is having its third Summit event today, and it’s bringing together some of the best minds in the world to solve problems of data center efficiency, server design, and more.

Most of its partner companies have a stake in the issue, either because they use massive server resources, because they design and sell hardware and chips, or because they are involved in creating the software that makes all this hardware more efficient.

“The momentum that has gathered behind the project – especially in the last six months — has been nothing short of amazing,” wrote Frank Frankovsky, Facebook’s hardware design guru, in a blog post this morning.

Frankovsky notes that the OCP now includes HP, AMD, Tencent, Salesforce, VMware, Canonical, Vantage, Alibaba, Supermicro, and Cloudscaling among its members, and that HP, Quanta, and Tencent have also joined the project’s Incubation Committee. This committee is responsible for reviewing proposals for official OCP support.

As far as new projects are concerned, Frankovksy said OCP has accepted proposals for a vanity-free storage server called “Knox”) and two high-efficiency motherboards, code-named “Roadrunner” and “Decathlete,” designed with the specific needs of financial services companies in mind.

OCP is also merging specs with Baidu and Tencent for its Open Rack design for servers.

Finally, the OCP project is doing what most open-source projects do these days: It’s launching support services for customers to easily and comfortably get started with its Open Rack designs. The OCP Solutions Provider program will allow companies to sell and use hardware based on OCP specs.

“Companies currently pursuing Solutions Provider status include Hyve, ZT Systems, and Avnet, as well as new business units from Quanta and Wistron (called QCT and Wiwynn, respectively) that have been launched to sell directly to consumers,” Frankovsky concluded.

The OCP got rolling about one year ago. At that time, Facebook’s vision of open-source hardware got started because Facebook itself was having trouble scaling its servers out in a way that made business sense.

“We looked at why things were done the way they were, and it always came down to legacy. Challenging legacies and starting from scratch was the most innovative thing we did in the project,” Facebook OCP lead Amir Michae told VentureBeat in an interview last fall.

“It’s natural in an environment where companies are trying to remain profitable to keep some pieces of innovation to themselves. But they also need to be able to share and engage with the community,” Michael concluded, encouraging other companies to get involved in the open-source hardware project.

Stay tuned for more coming up soon from today’s Open Compute Project Summit.

Also, if this is a story you find interesting, you should check out what Facebook is doing with Ringark, its mobile browser testing suite, and the W3C Core Mobile Web Platform Community Group.

Image courtesy of Jolie O’Dell

---

Big Data and Predictive/Real-time Analytics startups: Are you looking to jumpstart development & accelerate market traction? Sign up for the SAP Startup Focus program to receive technology, support, resources and community to help you develop new applications on SAP HANA, a cutting edge database platform. Get started here, and enter promo code “VB2013″ on the form.

---

But it didn’t mention one of the prime benefits of that shift. For every three servers that it used to have on Amazon Web Services, it can now get away with just one on the zCloud.

The zCloud can thus dramatically lower the company’s costs, Allan Leinwand, chief technology officer for infrastructure engineering at Zynga, told VentureBeat. That, in turn, gives Zynga some huge cost advantages over its rivals in the social game business, he said. Leinwand is sharing these details in a speech at the Cloud Connect conference in Santa Clara, Calif. today.

Zynga started out using its own servers. Then, in June 2009, it launched FarrmVille. The game was an enormous success and it forced Zynga to shift to Amazon Web Services, a public cloud-based infrastructure that allowed it to tap Amazon’s extra computing resources as necessary to handle the enormous demand for Zynga’s games.

FarmVille grew from zero to 10 million daily active users in its first six weeks and hit 25 million DAUs in its first five months. AWS gave Zynga huge flexibility, and it formally adopted a hybrid cloud strategy that mixed public and private clouds. Between 2009 and 2011, Zynga’s server capacity grew two orders of magnitude.

But starting last year, the company began heavily investing in the private zCloud. Yesterday, the San Francisco company disclosed that it spent $238 million on capital expenditures in 2011, up four-fold from the year before. And most of that was due to zCloud spending on server infrastructure that was precisely suited for Zynga’s mix of games. As it designed the zCloud, Zynga focused on the hardware and tried to eliminate every bottleneck in server components such as microprocessors, storage, and memory.

“We launched with AWS and then began to understand the stresses of the traffic and how we could manage it,” Leinwand said.

At the beginning of 2011, 20 percent of Zynga’s daily active users were hosted on the zCloud and 80 percent were on Amazon. Now that has been reversed. Over time, even more users will be hosted on the zCloud. Zynga was able to deploy the zCloud for its initial trial within six months. CityVille Hometown, a mobile game, was the first to be entirely hosted on zCloud. In November, CastleVille started on the zCloud, and the system was able to handle huge traffic growth when the game grew to five million users in six days. Zynga now has 240 million monthly active users.

Of course, shifting most of its user base to a private cloud carries its risks. Zynga now has to invest in its own data centers. If its games suddenly become unpopular and the user count drops, Zynga could get stuck with too much infrastructure. But the bet has paid off so far with steady growth in users and lower costs per user. Zynga can now deploy up to 1,000 servers in 24 hours. Zynga’s infrastructure is one of the largest uses of public and private clouds in the world.

Leinwand also said that Zynga has more control over the availability of its games. The company can now precisely understand the computing experience that players have, like how many seconds they have to wait for something to happen. Now it can measure everything. Leinwand said Zynga embraced the idea of “all-in availability,” which meant it would handle the load even if Amazon went down.

“We wanted to understand our availability and make sure that it was solid and always there,” he said.

Now that the zCloud is up and running, the company is collecting data on just how well all of the parts of the data center are working and what can be done to cut costs and improve the user experience. While Amazon is an all-purpose four-door sedan, Zynga’s custom-built zCloud is a sports car tuned to one purpose.

---

GamesBeat 2013 is our fifth annual conference on disruption in the video game market. You'll get 360-degree perspectives from top gaming executives, developers, and analysts on what’s to come in the industry. Our theme this year is “The Battle Royal.” Check out full event details here, and grab your early-bird tickets here!

---

Sept. 9 - 10, 2013
San Francisco, CA

Early Bird Tickets on Sale

Net Optics, a company that lets you listen in on your network activity, has acquired TripleLayer and nMetrics in an effort to create more easily-installed products in an anything but user-friendly industry.

“We adopted the use of a product kind of like Apple,” said Net Optics chief executive Bob Shaw in an interview. “You open it up, it’s simple to configure, and it just works.”

There are a few levels to a network. The first level is the actual hardware that exists in your data center. These data centers are filled with server racks that hold the physical equipment that makes your network run. But you need a way to listen to those networks, watch the processes happening inside the boxes. That’s where the second level comes in, tapping devices. Similar to what you think of with telephones, these boxes are connected to the network and essentially “listen in” on the conversations happening inside. They can tell where something is slowing down, if there’s a bottleneck, and can watch the network as information goes from one device to the next, a process that makes information vulnerable to attacks.

These taps exist across the market, but Net Optics decided its taps needed something more, it needed to address a new market.

Along with taps comes tools, such as McAfee, Palo Alto Networks, Imperva, and a number of other applications, that help secure and manage information coming through the network. These applications are expensive, and live separately from the taps and server hardware, as a software management layer. Installing them means taking down the whole system, integrating the tool and then booting the network back up. This can be more than costly for midsized companies, so apps like these address the enterprise market. They leave the little guy behind.

“Customers were saying that some of these tools are very expensive and companies who can afford these tools were only in the enterprise,” Net Optics founder Eldad Matityahu told VentureBeat. “So Bob and I decided to look at the middle market.”

The company turned to its distribution partner Triplelayer. Triplelayer, a small company in Australia, acted as a the gateway to countries in the Asia-Pacific and developed an eight-year relationship with Net Optics. The company’s sister, nMetrics, however, had less of a relationship with the company, but was developing interesting software, which Net Optics saw as an opportunity. The company scooped up nMetric’s technology to create a mid-layer between the applications and the tapping devices. Thus was born AppTaps.

AppTaps is a bit of software that lives on the taps hardware. When a tap is installed to the network, it comes preloaded with software that replaces security and monitoring apps on a very basic level. You don’t have to take the network offline to install AppTaps either. Thus, midsize companies are able to afford better coverage of their networks. The coverage with AppTaps is better than just putting in a listening device. In addition to watching the network, AppTaps can give you analytics about what’s happening, sends alerts when something goes wrong, and sends push notifications to your phone and email.

All of Net Optics’ high profile clients such as Google, eBay, IBM, Oracle and AT&T are installing AppTaps. Indeed, AT&T wants 1,000 AppTaps installed at its endpoints — the towers and individual stores — so that it can monitor network activity from computers it usually doesn’t have access to. Because of its success, Net Optics decided to assimilate TripleLayer and nMetrics officially into the company.

The two companies are “desks across each other in a room,” according to Matityahu, but pack a lot of power. They total seven employees and two co-founders, one for each sister company. TripleLayer will now handle all of distribution for Net Optics’ Asia-Pacific shipping, and nMetrics will continue to write code for the AppTaps product.

Net Optics has not revealed the purchase price.

“HP has secretly contracted with Intel to keep churning out Itaniums so that HP can maintain the appearance that a dead microprocessor is still alive,” according to a filing obtained by All Things D. “The whole thing is a remake of Weekend at Bernie’s,” the filing states.

Oracle and HP have been in a contentious legal battle since June, when HP sued Oracle for ending software support for the Intel Itanium chip. HP is the only major vendor that sells servers with the Itanium chip, but the most popular servers (and software for those servers) run on Intel’s x86-based chips.

HP claims Oracle ended its Itanium support because Oracle acquired Sun Micrososystems and its server business, which use the x86-based chips. HP also has suggested that Oracle wants to convince current and potential Itanium server customers that they should only be interested in x86 servers.

Intel said in a statement, “We are not a party in this litigation and will not comment on litigation where we are not a party. And, we never comment on confidential agreements between us and other companies.”

And HP’s response to the filing reads:

This filing is just the latest in [Oracle's] ongoing campaign to shore up its failing Sun server business and starve thousands of existing Itanium customers who rely on their Itanium processors for mission-critical activities.

As Oracle well knows, HP and Intel have a contractual commitment to continue to sell mission-critical Itanium processers to our customers through the next two generations of microprocessors, thus ensuring the availability of Itanium through at least the end of the decade. HP is resolved to enforcing Oracle’s commitments to HP and our shared customers and will continue to take actions to protect its customers’ best interests. It is time for Oracle to quit pursuing baseless accusations and honor its commitments to HP and to our shared customers in a timely manner.

Your Facebook profile doesn’t exist on your computer or in some nebulous cloud called “the Internet.”

It’s stored deep in the brick-and-mortar walls of real-world fortresses. It comes to life as electricity flows through wires that connect tens of thousands of servers to the grid.

And for some users some of the time, it lives among the wind and scrub brush of central Oregon, where Facebook has erected its first fully functioning data center in a town called Prineville.

The still-young company leases equipment and facilities at various locations, but the Prineville center is something special. Facebook designed and built this place from the ground up. More interestingly, it’s shared its customized hardware designs and super-efficient operational specs with anyone who wants to see them.

Yesterday, we spent the afternoon poking our nose around Facebook’s Prineville data center. We’ll have a longer video tour of the place posted soon, but we wanted to share the images from the trip as soon as possible.

Enjoy the data center porn, and be sure to read up on why Facebook thought open-source hardware was so important in the first place.

Ushering in the era of low-power servers, Austin, Texas-based Calxeda is today announcing its EnergyCore ARM-based processor, the first ever chip capable of running an entire server at a mere 5 watts.

The EnergyCore server-on-a-chip uses 90 percent less power (just 1.5 watts while idle), takes up 90 percent less space, and is half as expensive as traditional server solutions, according to the company. Since it’s based on ARM technology, Calxeda’s chips are taking a cue from the low-power, yet highly capable, ARM processors used in smartphones and tablets.

“ARM is to processors what Linux is to operating systems,” Calxeda VP of marketing Karl Freund told VentureBeat in an interview yesterday, referring to the way companies can build innovative technologies on top of ARM’s original designs. The EnergyCore chips are based on ARM’s quad-core Cortex A9, and they run at speeds between 1.1 gigahertz and 1.4 Ghz. But the company also added in an 80-gigabit fabric switch, which will allow for high data throughput, as well as an energy management engine.

The complete EnergyCore server node also includes 4 gigabytes of RAM. Calxeda’s chips are 32-bit, but the company says it will be ready to jump to 64-bit chips once ARM’s designs are complete.

Calxeda says its chips are best suited for target applications like storage and file serving, or web apps. You won’t see much of an advantage using EnergyCore servers for heavy duty video encoding, but for most other server uses it’s an ideal balance between low-energy usage and a decent amount of computing power. The company competes directly with Intel’s Atom chips and firms like SeaMicro who are building Atom-based servers.

Calxeda already has a launch partner in HP, which today announced its new server lineup based using the company’s technology. “A single rack of HP’s Calxeda servers delivers the throughput of some 700 traditional servers and dramatically simplifies the infrastructure needed to hook them all together and manage the cluster,” said Calxeda CEO and co-founder Barry Evans in a statement today.

Calxeda, formerly known as Smooth-Stone, raised $48 million in funding about 14 months ago from ARM, Advanced Technology Investment Company (ATIC), Battery Ventures, Flybridge Capital Partners and Highland Capital Partners.

The Open Compute Foundation, announced today, will allow anyone to access the designs and specifications for Facebook’s homebrewed and highly efficient data center hardware and will provide structure for the project.

To join, hardware designers and hackers need to sign an agreement on the Foundation’s site, find an aspect to work on and sign an open licensing agreement to do that work. Designs will be voted into the official project based on merit.

The goals are to make data center energy usage more efficient, to make server repairs faster and fewer and to figure out better ways to serve data at massive scale, e.g., billions of users, what is known as “human scale.”

“This is not going to be a community that’s going to write wish-lists,” said Frank Frankovsky, one of the founders of the Open Compute project at a summit in New York City this morning. “We’re publishing not only specifications but also source files.”

“When we first launched this project back in April, people thought we were crazy,” said Frankovsky.

Now, however, Facebook’s groundbreaking work in data center efficiency and server redesign will be open for all to hack on and improve. As an example, Frankovsky mentioned Facebook’s newest data center in Sweden, a hydropowered facility that he said is “our greenest yet.” That design will be open sourced, its specs published and available for anyone to manipulate and possibly enhance.

Right now, you can go to the Foundation website, login with your Facebook and Github accounts, and start checking out the components that make up Facebook’s data centers, from the motherboards to the chassis — even Facebook’s completely redesigned power supplies.

“Let’s build this together,” said Frankovsky. “If we start sharing our ideas, the pace of innovation is going to increase rapidly. … Let’s start focusing on the environment and the efficiency of this.”

Asus, Intel and AMD are involved as partners and will be publishing specs and source files, which hackers will then be allowed to modify and submit back to the project. Dell is also a major partner for the project.

Facebook’s Open Compute project has inspired the interest of the hardware hacker and open-source communities, individuals and groups that Facebook said “are passionate about making strong technical contributions to defining and delivering the most efficient server, storage and data center designs.”

Open source is a huge part of Facebook’s culture. Pretty much since its inception, the company has both used and created or contributed to open-source software projects. In a recent conversation with Facebook open-source software lead David Recordon, we talked about how open-source is simply part of Facebook’s DNA in a way that’s rare for a Silicon Valley startup.

“I think it’s pretty clear there’s no question about whether companies should be using open-source software or not,” said Recordon to VentureBeat back in August. “That was answered over the past decade. The question now is about open hardware. Many of the things that we have today for open-source software we don’t have for hardware and standards.”

Amir Michel, who heads up Facebook’s open-source hardware efforts, also told us in an interview on Facebook’s data center redesign, “A lot of the tools aren’t there yet,” said Michael. “If someone wants to make a change to one of our circuit boards, it takes hundreds of thousands of dollars to get that package. The average hacker doesn’t have that. Most of the contributions so far come from other large companies. We’re hoping to change that in the future so a guy in his garage can design a motherboard.”

Stay tuned for more news on open-source hardware as the Open Compute Foundation begins accepting projects from new contributors outside Facebook.

Check out DevBeat, VentureBeat’s brand new channel specifically for developers. The channel will break relevant news and provide insightful commentary aimed to assist developers. DevBeat is sponsored by the Intel AppUp developer program.

---

Big Data and Predictive/Real-time Analytics startups: Are you looking to jumpstart development & accelerate market traction? Sign up for the SAP Startup Focus program to receive technology, support, resources and community to help you develop new applications on SAP HANA, a cutting edge database platform. Get started here, and enter promo code “VB2013″ on the form.

---

Skaugen spoke because his boss Paul Otellini, chief executive of Intel, was sick. He said that Moore’s Law, the prediction made by Intel chairman emeritus Gordon Moore back in 1965 that the number of transistors on a chip would double every two years, is expected to hold up in the next few years.

Because chip technology continues to advance, the servers that use those chips can handle more and more traffic in data centers. As a result, users are able to upload 48 hours of video to YouTube every minute. Users can also send tweets about 200 million times a day. The internet now supports more than 4 billion connected devices today, and about 7.5 billion photos are uploaded to Facebook each month.

In the past, the number of processors in data centers grew two-fold in ten years, from 2000 to 2010. Skaugen believes we’ll see another doubling in five years. The average price per server has dropped in that time from $58,000 to $3,800.

Virtualization, or software that helps spread processing loads across more processors, will be so popular by 2015 that about three fourths of the time you will get a virtual server when you order a new server. Skaugen said we generated 245 exabytes of data last year and we are spending $450 billion on data centers each year. During 2011, we’ll generate 300 exabytes of data.

The data and the processors will be needed in the future to make increasingly accurate weather predictions. Skaugen said that servers will become 125 times more powerful by 2018 and they will consume about twice as much power. At that point, weather forecasters will be able to predict the landfall locations for hurricanes within a 100-mile radius. Eventually, forecasters should be able to predict the exact zip code where a hurricane will hit, days ahead of time.

While there are concerns that power limits will curtail Moore’s Law, Skaugen said, “Moore’s Law is alive and well.” He said that Intel is working on several generations of chips now that will enable faster processing for the next few years that stays on the pace demanded by Moore’s Law.

“The data center market continues to grow at 8 to 12 percent per year. However, it is becoming constrained by the availability of power, especially in high density areas where space is a limitation,” said Tarak Mehta, Head of ABB’s Low Voltage Products division in a statement.

Managing large scale data centers is a cumbersome task when combined with powering multiple computers on one physical server, cooling costs and more. Power Assure attempts to tackle enterprise- to government-size data center management issues, such as dealing with large amounts of collected data, including information on a server’s own performance. The company provides software to track and analyze this data, allowing data center managers to make better decisions on how to run their servers more efficiently as well as how to optimize usage to save power and free up server space.

The funding will be used to increase marketing and sales efforts. The company will also invest in product enhancements for its energy management solutions.

Power Assure is headquartered in Santa Clara, Calif. The company has raised $28.75 in total from investors ABB Technology Ventures, Draper Fisher Jurvetson, Good Energies, Point Judith Capital. The United States Department of Energy also provided Power Assure with a $5 million match-funding grant.

[Photo courtesy of Power Assure]

Sept. 9 - 10, 2013
San Francisco, CA

Early Bird Tickets on Sale

Major cloud computing services provider Xtium has raised $11.5 million to expand its reach in helping mid-size companies with private cloud computing, virtual hosting and virtual disaster recovery.

Xtium is one of the leading infrastructure-as-a-service (IaaS) providers in the U.S., but it has to compete with major players like Amazon and IBM. The company offers as pay-as-you-grow model to businesses that allows for more flexibility when companies need back-end solutions like virtualization, servers and storage. Xtium said it often attracts customers with its “cloud disaster recovery service” and then eases them into further services after they like what they see.

The company plans to use its new capital for expansion, hiring more employees and further investing in its cloud technologies.

“Xtium is having another strong growth year and we are excited to continue building for aggressive execution,” said Peter Ritz, president of Xtium, in a statement. “Scores of mid-market enterprise customers have chosen Xtium to migrate and manage their cloud and network, as we provide everything they need to plan, migrate, and manage their systems and network at a significantly lower cost and higher service level than their current model.”

Valley Forge, Penn.-based Xtium was founded in 2004 and started managing cloud customers in 2007. Its first round of funding was led by OpenView Venture Partners, a Boston-based firm that has backed startups such as Mashery, Monetate, nextdocs, Skytap, Balihoo and Instructure.

It’s one thing to open-source the code for your app — that’s a simple matter of mashing a button on Github. But how do you really open-source hardware?

Think about that: Facebook committed to open-sourcing the infrastructure of its data centers through the Open Compute Project, which launched back in April. But there’s more to maintaining an open-source project than just releasing data into the wild. You also have to accept contributions from other members of the community.

So how do you accept a patch for a motherboard? Or an improvement to a power supply?

This was just one of many challenges facing Amir Michael and the rest of Facebook’s open-source hardware team as they began redesigning the company’s servers and data centers. And to be frank, it wasn’t even the most challenging problem they’ve faced so far.

How Open Compute began

Michael, a former Googler, told VentureBeat that when he first came to Facebook, “I knew a lot about servers and data centers.” Not only did he understand the architecture of a network of servers; he even had hands-on experience as a data center tech, where he often worked until his hands were raw from repairing downed machines and replacing faulty components.

At the time Michael first came on board at Facebook, he said, “The way Facebook was scaling was tremendous. We were buying servers from HP and Dell, leasing server space from Data Realty Trust.”

But Michael had an inkling that one of the most significant companies of the decade might not actually have been handling its data in the smartest, most efficient way. “I did a little analysis,” he said. “I went to NewEgg.com and put together an equivalent server, and it was about the same price, even though we were buying in these huge volumes.

“The business model didn’t make sense.”

After realizing that Facebook wasn’t doing itself any favors buying stock servers in huge quantities, Michael started investigating how the servers were cooled and powered. “I realized there was a lot of inefficiency there, too,” he said. “We looked at how to improve it. With optimizing the data center and ignoring the servers, you get some efficiencies, and you get some efficiencies by optimizing the servers and ignoring the data centers. But you get the biggest benefits if you optimize both.”

In a nutshell, that’s how the Open Compute Project was born.

At the outset, Michael and the Facebook team tried to work with their existing hardware providers. “The vendors’ responses to the changes we wanted to make were lukewarm,” Michael said. “They offered to do a bunch of other things that weren’t too useful for us. They wanted us to buy what their other customers were using, but those modified machines weren’t as extreme as the customizations we were considering.”

Redesigning the server

From that point forward, Michael, Facebook’s manager for hardware design, started tearing apart every assumption about how servers were supposed to be built.

“We looked at why things were done the way they were, and it always came down to legacy. Challenging legacies and starting from scratch was the most innovative thing we did in the project,” he said.

For example, in Facebook’s new server design, the way power is delivered to the microprocessor is entirely different. The team took out transformations and distribution mechanisms and changed the power supply itself. Even the power cords and power strips have been entirely re-engineered, and the servers themselves were designed to be built and maintained without any tools.

In fact, Michael said the serviceability of the server was one of the team’s most important innovations. “When you have tens of thousands of servers, they break on an hourly basis. The hard drives fail, the memory fails. Our data center technicians are responsible for maintaining the servers. They spend their whole day installing new cabinets, new hard drives, etc. We wanted to make their jobs as easy as possible and a lot more efficient. We didn’t require any tools to assemble the servers, and most components are two to 10 times faster for basic service functions than on an average server.”

To test out this aspect of efficiency, Facebook had a prototype build party, which goes down in our book as one of the nerdiest ways to have fun on a Saturday night. “We let a bunch of engineers build the servers, we had pizza and beer, and we had a competition to see who could build a server the fastest,” said Michael. “A data center tech got it built in eight minutes.”

Facing resistance

When Michael was done redesigning the most fundamental aspects of the server, however, he didn’t get an initial enthusiastic response from a few key audiences. Engineers at Facebook who had to do work on the servers were “skeptical,” he said, and even the new vendors were “hesitant.” Facebook’s management took some convincing, as well.

“In general, the resistance to change, getting people to accept a new architecture, was our biggest challenge,” he said. “Getting people to be open to trying something new was hard.”

That’s especially true for big hardware changes. Making radical software changes is, by contrast, cheap and easy. “With hardware,” said Michael, “you need a lab, new hires, prototypes. It requires several million dollars worth of investment. To their credit, Facebook management’s willingness to invest in this fringe project speaks to their ability to take big risks and allow for innovation to occur.”

Those big risks involved trip after trip to Taiwan to work with new manufacturers, bringing a mechanical engineer in-house, and drafting between 50 and 60 pages of specs for the new servers. “Doing design on a white board is one thing, but figuring out the details is where you can stumble, said Michael, “especially when everything you’re doing is customized and entirely new from the ground up.”

Open-sourcing hardware

Finally, the Facebook team is still trying to figure out how to make the Open Compute Project truly open source by accepting contributions from the hardware hacker community.

“A lot of the tools aren’t there yet,” said Michael. “If someone wants to make a change to one of our circuit boards, it takes hundreds of thousands of dollars to get that package. The average hacker doesn’t have that. Most of the contributions so far come from other large companies. We’re hoping to change that in the future so a guy in his garage can design a motherboard.”

Michael continued to say that with the right software, that garage hacker could be making contributions as innovative as anything coming from a lab at HP or Dell. Currently, even the software used to design hardware is prohibitively expensive. But this is code — invisible, intangible ones and zeroes — and there’s no reason it shouldn’t be free.

Facebook wants to work with software vendors on free licenses for Open Compute Project contributors. The company is also considering working with other corporations and organizations (such as governments and large universities, which have similar computing needs) to create new, open-source software programs for hardware design.

Another prohibitive aspect is prototype creation. A typical prototype server might cost between five and 10 times more to build than a production server, so even garage hackers might need to get some kind of financial backing for those projects.

The philosophy of open-source at Facebook

We asked Michael if he had any ideological qualms about being an open-source guy at a proprietary software company. “As the guy who builds the infrastructure, I’m disconnected from the software that runs the site. It’s not a dilemma I experience on a daily basis,” he said.

But he continued, “It’s natural in an environment where companies are trying to remain profitable to keep some pieces of innovation to themselves. But they also need to be able to share and engage with the community. If you think about our business model, it’s about providing a valuable service to our users. The infrastructure we use to do that wasn’t a key piece of the business model. Our advantage is the product, not the servers. It’s not a core piece of IP.”

Also, Michael said, “Engineers are social beings, too, and they like being able to talk about the things they’re passionate about. And when you share information, you get benefits. You get feedback from other people about better, cheaper ways to do things.

“If you look at how Facebook was built, it uses a lot of open-source software. We’ve contributed back a lot in the software world, but we haven’t contribued back to the hardware world yet. No one has. But if we do that, maybe other companies can use the same kind of infrastructure. They don’t have to waste energy, and they don’t have to go through the same development process we did.”

Sharing information with universities has been particularly fruitful, Michael told us. “They have interesting solutions, but they don’t have enough data about real-world problems. They don’t know how industries operate. So by sharing information about our workloads and configurations, we get a lot of interest from universities.”

“Then there’s the environmental impact,” he said. “If we share these best practices, we’re hoping that other people can adopt it and have an impact on the environment as well.”

Check out DevBeat, VentureBeat’s brand new channel specifically for developers. The channel will break relevant news and provide insightful commentary aimed to assist developers. DevBeat is sponsored by the Intel AppUp developer program.

---

Big Data and Predictive/Real-time Analytics startups: Are you looking to jumpstart development & accelerate market traction? Sign up for the SAP Startup Focus program to receive technology, support, resources and community to help you develop new applications on SAP HANA, a cutting edge database platform. Get started here, and enter promo code “VB2013″ on the form.

---

The company provides a single sign-on for access to servers in addition to management solutions. These solutions include the ability to turn off potentially compromised machines, restrict user privileges, and monitor all activity on the servers.

“A lot of breaches come from insiders who have the keys to the kingdom. We eliminate the keys to the kingdom,” chief executive Tom Kemp told VentureBeat in an interview.

Kemp explained that having multiple sign-ons for various applications on the server is a vulnerability. For instance, firing an employee requires the disabling of his user names and passwords. But this leaves room for access points to fall through the cracks. You may have removed their Google Apps access, but what about the Salesforce account you forgot about?

If there is only one sign-on, however, you can easily retract server access from Centrify’s hub. Okta, which also recently announced a round of funding, is a direct competitor to Centrify’s cloud security offerings.

But what Centrify is most concerned about is how many different devices can now access server information.

“The data center of the future is very heterogeneous and very hybrid,” said Kemp.

He means smartphones and tablets are quickly entering the enterprise as viable ways to distribute data. But these devices are easily compromised. Not everyone password-protects their mobile devices, and they are often forgotten on a train seat or in a restaurant. This pokes an immediate hole in the security of that company’s server infrastructure — cloud or on-premise.

The funding was led by Index Ventures.

Centrify is using some of the funds to build a product combating this vulnerability. The product will allow server managers to deactivate any mobile device associated with a company, in the same way that they can deactivate computers and servers.

While simple, this could bring some peace to worried managers as more and more devices store proprietary information in the cloud.

But while a single sign-on is a good way to manage security, a watchful eye still needs to be placed on existing user activity.

“We do have an auditing solution that provides, in effect, a security camera on the servers and logs all activity done on that server,” said Kemp.

That security camera can be set up to send alerts to a manger and can even be programmed to shut down all systems when a potential security threat is identified. This latter solution could be more effort than it’s worth given the possibility of false alarms.

Another flaw is the inability to detect security threats during work hours. Some threats may come at night and are easily detected. But what about the ones that occur during the normal pace of work? Many servers are exporting and importing data regularly. And if what Kemp says is true, and many breaches come from the inside, it is even harder to detect usual operations versus threats.

According to Kemp, Centrify cannot identify those breaches at this time.

Centrify will also use the funding to expand internationally. Kemp is particularly excited about leveraging Index Ventures’ international presence and believes the firm will play a role in Centrify’s expansion.

Currently, the company has over 3,500 customers including Research in Motion, Pfizer, and the U.S. Army. It has 150 employees, headquartered in Sunnyvale, Calif. Thus far, Centrify has accumulated $52 million in funding from Index Ventures, Mayfield Fund, Accel Partners, INVESCO Private Capital and Sigma Partners.

Facebook, one of the biggest consumers of computing technology, is announcing today that its study showed Tilera’s TilePro processor has the best performance per watt in the industry. The study says that Tilera-based servers can provide three times the performance for a given amount of energy as Intel’s x86-based servers. The ratio is four times better than Advanced Micro Devices server chips.

“This is the first third-party validation of our technology,” said Ihab Bishara, director of cloud computing products at Tilera, in an interview.

The finding is important because it’s critical for data center operators to pick the right configuration of chips and hardware for their operations that involve serving millions of web pages to users on a daily basis. Companies such as Twitter, Zynga and Flickr — all of which have tens of millions of users or more — are likely to pay close attention.

Facebook ran tests with x86-based and Tilera-based servers using a common data center application dubbed Memcached. The results of the study will be released at the Second International Green Computing Conference in Orlando, Fla. Facebook tested a Tilera-based server, the S2Q, which is built by Quanta Computer. Quanta supplies servers for Facebook’s open compute platform (data centers). The S2Q has eight Tilera chips in it. Each chip is a TilePro64 (pictured) with 64 cores, or computing brains, on a single chip. Both the x86 and Tilera servers could process transactions in less than a millisecond and had comparable delay times. But Tilera’s processors performed more transactions per second at the required latency level.

The TilePro64 has 67-percent higher throughput, a measure of how fast data can be processed in a second, compared to the x86 servers. Tilera does better because it has a way to interconnect its cores on a chip so that the cores can work in parallel without consuming a lot of power, which has become the highest cost in a data center, said Bishara.

Tilera says it is not allowed to say whether or not Facebook uses its chips in servers in its data center, which has to support more than 750 million users as they access their Facebook pages. But Facebook plans to run the same study on Tilera’s new 64-bit Gx3000 series chip, which was announced in June. The Gx3000 will be available in samples later this month and can handle 10 times the performance per watt compared to Intel’s processors.

Besides Facebook, other popular Memcached users include: Zynga, LiveJournal, Wikipedia, Flickr, Bebo, Twitter, Typepad, Yellowbot, YouTube, Digg, WordPress, Craigslist and Mixi. Memcached is used to provide fast response times by keeping data stored in memory rather than on slower drives.

Tilera was founded in 2004 and has fewer than 100 employees. To date, the company has raised $109 million. Rivals include Cavium Networks, RMI, Netlogic, Freescale, Intel and others. Tilera’s investors include Artis Capital, West Summit, Cisco, Samsung, Walden International, Bessemer Venture Partners, Columbia Capital, Broadcom, NTT Finance, VentureTech Alliance and Quanta Computer.

SeaMicro shook up the world of data center computing by making energy-efficient servers built around Intel’s low-power Atom chips, which are usually used in netbooks and laptops.

“This is more computing power in a given space than anyone has ever built for an Intel system,” said Andrew Feldman, chief executive of SeaMicro, in an interview. “In the past nine months, we have brought a drumbeat of innovation to the server industry.”

The advantage of Sunnyvale, Calif.-based SeaMicro’s small and power efficient computers is that enterprises can now shove a lot more computing power into a given amount of space and use a lot less electrical power, thereby cutting costs such as electricity bills dramatically.

Sales of previous systems have been doubling each quarter, Feldman said. If that pace continues, Feldman said SeaMicro could become the fastest-growing system company in Silicon Valley history. SeaMicro’s customers include France Telecom, Skype, Rogers Wireless, Mozilla, eHarmony, and China Netcom BB. Hundreds of millions of internet users traverse SeaMicro’s hardware daily.

Intel has been improving its server microprocessors by making them more power efficient. But the microprocessor only accounts for a third of the power consumption in a server. SeaMicro’s innovation lies in how it attacks the remaining two-thirds of the power consumption problem. It does so by combining a lot of the extraneous chips into a single, more-efficient custom chip, said Feldman said.

SeaMicro can put 384 Intel Atom dual-core processors (for a total of 64-bit 768 cores) in a 10-rack system, which is just 17.5 inches high. The new SM10000-64HD is a 20 percent improvement over SeaMicro’s previous server line and a 150 percent improvement on its compute density record, or the amount of computing power in a given space.

This single machine can replace rival systems with a bunch of equipment: 60 traditional servers, four rack switches, four terminal servers, and a load balancing server. It uses a quarter of the power and a sixth of the space.

SeaMicro can put an entire server on a motherboard (pictured below) that is 5 inches by 11 inches. Since the hardware is Intel-based, customers don’t have to change their software at all.

Intel no doubt likes the fact that SeaMicro can work such wonders with Intel Atom N570 chips while maintaining software compatibility.

“SeaMicro continues to drive innovation in the microserver market,” said Jason Waxman, general manager of Intel’s Data Center Group, who noted that SeaMicro can get the equivalent of 3,000 processors running at a combined total of 5,100 gigahertz.

But Feldman notes that SeaMicro could just as easily use ARM processors in its servers, if customers wanted that. SeaMicro’s previous SM10000-64 server had 256 dual-core Atom processors for a total of 512 cores in a 10-rack unit.

Feldman acknowledges the SM10000-64HD is a pretty nerdy name for the company’s latest server. I suggest they call it the Atom Bomb.

Zeus Kerravala, an analyst at the Yankee Group, a market analyst firm, said that SeaMicro has delivered three servers in nine months, each time setting the record for energy efficiency and compute density. The list price is $237,000 for the new server, which has been shipping for 10 days. Feldman said previous units have sold millions of dollars worth.

SeaMicro was founded in 2007 and is backed with $60 million in funding by Khosla Ventures, Draper Fisher Jurvetson, Crosslink Capital, as well as a number of strategic investors. It also won grants from the U.S. Departement of Energy ($9.3 million) and the state of California ($250,000). SeaMicro has 80 employees. Most recently, SeaMicro raised $20 million.

Here’s what we said in past stories about how SeaMicro’s servers work:

SeaMicro also attacked the power consumption with a very clever trick known as virtualization.

Today, virtualization is frequently used with servers. It is a layer of software that rests between an application and the servers that it runs on. If an application needs only two servers, the virtualization software finds two available servers to run the application. If the application gets busy and needs 10 servers, the virtualization software finds 10 available servers to do the job. The application is no longer tied to specific servers; the virtualization software frees up the overall system and gets more utilization out of the available servers.

SeaMicro did the same thing, but it applied the concept of virtualization to the inside of a server. Feldman designed custom chips that could take the tasks that were handled by everything beyond the Intel microprocessor and its chip set. The custom chips virtualize all of those other components so that it finds the resource when it’s needed. It essentially tricks the microprocessor into thinking that the rest of the system is there when it needs it.

SeaMicro virtualized a lot of functions that took up a lot of space inside each server in a rack. It also did the same with functions such as storage, networking, server management and load balancing. Full told, SeaMicro eliminates 90 percent of the components from a system board. SeaMicro calls this CPU/IO virtualization. With it, SeaMicro shrinks the size of the system board from a pizza box to the size of a credit card.

By boiling down the rest of the system into a couple of chips, SeaMicro can get rid of a lot of the components in a system, thereby getting rid of space, cost, and power consumption.

JouleX

The company creates software that gives other businesses a way to monitor the energy usage of every device attached to a computer network. That business can then write specific rules for how it wants to conserve energy or eliminate carbon emissions — like powering down lights or certain data centers when they aren’t in use. Businesses can also run programs that simulate how those devices will operate under energy-conserving situations.

The software detects personal computers, Mac computers, voice-over-Internet phones, printers, network switches and routers, heating and cooling units, storage servers and just about every other kind of device that’s attached to a local network. It then creates a profile for each device that details how much energy it is using and how much carbon it is emitting. It doesn’t require any extra hardware to create each profile.

JouleX focuses on the enterprise space, where companies regularly use massive data centers to run their programs. That’s because most companies specialize in some form of cloud computing, which run processes like large-scale calculations or video compression on powerful servers and then deliver the results of those processes to a computer connected through a network or through the Internet.

Cloud computing lets companies reduce their costs because they don’t have to buy powerful computers for every person in the company. They can focus on purchasing powerful servers instead and then create virtual desktops for each employee on weaker, cheaper computers. But those data centers require a lot of electricity and aren’t really all that green. That gives companies like JouleX a business opportunity by giving companies like Cisco a chance to reduce their electricity costs.

Existing investors Target Partners and TechOperators join Sigma Partners and Flybridge Capital Partners as investors in the most recent funding round. The company has around 100 customers that include the likes of Cisco and Intel, according to its website.