The boom in Internet videos transfers — from movie downloads to amateur video sharing — is causing pain in the Internet backbone.

Now a Palo Alto start-up called Itiva has popped up, saying it has found the magic pill: It says it can do something that no other start-up has offered so far: (1) It can cheapen the cost of video transmissions for big movie or video producers, and thus match the promise of the peer-to-peer technologies out there, and at the same time (2) lower the cost for the Internet service providers.

Here is the Mercury News story we ran today about Itiva and some of the other start-ups out there helping with video transfer technology.

Itiva works like this:

When a user downloads a large video file, say the latest SF Giants game from, the individual pieces of the file are saved in that cache of the Internet service provider’s network. Once there, those video pieces are latently ready so that if any other user wants to download the same game, they can do so — and they don’t have to go back to the site. It saves MLB cash, and it also saves cash for Comcast, if that is the user’s Internet service provider. Below is a diagram (click to enlarge).

In our Merc story (link above), we explain Itiva’s positioning in relation to peer to peer folks like BitTorrent and Pando, and as usual we didn’t have much space to delve into the subtleties.

For example, while Itiva’s technology may work for popular broadcasts, such as baseball games, it won’t work very well for niche programs downloaded by smaller audiences — where there isn’t enough streaming to cache it all.

Michel Billard

And the technology, while being tested by Dreamworks and others, is still too early to assess. The analysts we talked with still haven’t kicked the tires properly on Itiva. So it is still he-said-she-said. For example, ATR analyst Albert Lin says Itiva cache process may be vulnerable to access by others, and thus something content owners won’t like. “There’s no guarantee a particular piece of content is able to be blocked or controlled,” says Lin. Itiva does have security guarantees, Itiva President Michel Billard responds. “It’s encrypted. No one can access it without making a request to our server.”

Even if Itiva does make things secure, it faces a challenge from BitTorrent, which is busy developing its own cache capability. We reached Ashwin Navin, president of BitTorrent, while he was in London last month, managing trials with the UK’s largest cable company, NTL. NTL has acquired cache technology from a company called Cachelogic, specifically to manage BitTorrent’s traffic, he said. For now, Ashwin says he is confident BitTorrent’s lead will hold. BitTorrent has between three to five million users a year, and traffic is up 50 percent or more compared to last year, he said.

Still, Itiva’s Billard points out that NTL is still having buy the cache — which represents a cost — something it wouldn’t have to do with Itiva’s technology. So, the question is whether Itiva can really deliver a service that is radically better for all sides, so that it can break into the lead of a player like BitTorrent. The other thing to keep in mind is that BitTorrent’s file sharing happens across service provider networks, and there’s a cost (for service providers) each time a file transfer happens across those different networks. Itiva’s advantage is that it operates within a network, and so lowers cost. But operating within a single networks can also be a disadvantage, since it means the chances of finding a particular video within a single network’s cache is lower than if you have the entire Web available.

Finally, another point is that Internet service providers might decide they like certain peer-to-peer technologies, even if it is hogging up their networks. AMR’s Lin says Pando’s user-friendly technology could help some service providers differentiate themselves from competitors. By offering Pando on a set-top box, giving users the ability to receive large movie files, for example, a service providers might convince customers of its value – a plus that may outweigh any costs caused by Pando’s peer-to-peer bandwidth usage.

Others are skeptical of peer-to-peer. Says Insight analyst Robert Rosenberg: “”The stakes are very high, and what you want more than anything else is reliability, and that’s not in the peer-to-peer environment.”

So on Itiva, we will see. Itiva has raised $7 million from founders and angel, and it has engaged Kaufman Brothers LP to help it look for another $10 million. If Itiva’s trials with DreamWorks are going well, and it shows those results to potential investors, raising cash should be pretty damned easy in this environment!

Update: Stu Phillips, an investor in Cacheflow (now BlueCoat Systems) chimes in here.