x
Optical/IP

Startups See More Cache in Carrier P2P

Peer-to-peer traffic is nothing new, and how to deal with it has been the subject of some debate -- and opportunity. With some 40 percent of all network traffic stemming from P2P applications, some carriers have turned to deep packet inspection (DPI) and traffic managing solutions, but a few startups are betting that carriers may need more to ease their pain. (See P2P Takes Over.)

That's where P2P caching companies like Oversi Networks Ltd. and PeerApp Ltd. come in.

Oversi and PeerApp sell equipment that sits in the carrier network and caches content from recognized P2P protocols. When more than one user downloads a P2P file on the same network, the file is served up from the local cache rather than from a remote server or peer. The end result is less traffic passing through carrier peering points and faster delivery of the P2P content.

Israel-based Oversi has been operating since 2004 and offers a P2P caching solution, online network storage, and a content distribution platform. Its Overcache P2P caching series of products has a 90 percent byte hit ratio and claims it can scale to support traffic of up to 100 Gbit/s.

Newton, Mass.-based PeerApp was founded in 2004 and has its R&D operations in Tel Aviv. The company's Ultraband family of products promises up to 2 Gbit/s capacity with a P2P byte hit ratio of up to 80 percent. The Ultraband products, the company notes, are completely transparent to the network.

But while Oversi's P2P caching product only captures downstream traffic, PeerApp's provides upstream and downstream caching, which it says can help to mitigate traffic congestion. PeerApp also received a patent for its technology earlier this year. (See PeerApp Awarded Patent.)

PeerApp has financial backing from Pilot House Ventures , Cedar Fund , and Evergreen Venture Capital . Oversi, meanwhile, just received funding from Cisco Systems Inc. (Nasdaq: CSCO) and has previously received funds from investors StageOne Ventures and TempoPark Fund. (See Cisco Invests in P2P Video Vendor.)

Both companies already have customers for their products, although neither has named a major buyer of P2P caching devices. So far, the customers for this kind of gear are second- or third-tier providers in developing nations, where transport costs are high.

PeerApp claims 12 service provider customers, mostly in Asia and South America. Oversi says it has 20 customers worldwide, most of which are also in Asia, with some customers in Europe and South America.

But both startups claim there is increased interest in their technology by Tier 1 providers. That's due in part to the amount of legitimate content being delivered using P2P protocols.

"In the U.S., peer-to-peer was a dirty word," says PeerApp's VP of business development, Frank Childs. "But now we're seeing business models leveraging this," which he says is causing U.S. carriers to pay attention.

Some carriers may be put off by the idea of caching content they can't guarantee was downloaded legally. Oversi and PeerApp both claim compliance with the DMCA, but both admit that their caching products currently don't recognize what kinds of content are being served. "We intentionally do not know" what content is being cached on PeerApp's devices, Childs says.

But with the increase in the number of ad-based Internet video providers leveraging P2P technologies to deliver content, differentiating among different types of P2P content could be of value to service providers.

"Service providers want to deliver content without being a dumb pipe," says Eitan Efron, Oversi's co-founder and VP of marketing. In the future, he believes, P2P caching products will be able to differentiate among premium, non-premium, user-generated, and illegal content, which could allow carriers to charge more money for certain types of content.

As a result, Oversi and PeerApp are looking to add protocols from P2P-based content providers. Oversi and PeerApp have for the most part reverse-engineered P2P protocols used by eDonkey and Gnutella, but they hope to partner with legit services like Joost and Babelgum.

While Oversi and PeerApp are just picking up steam in the P2P caching business, it's worth noting that at least one player in the segment recently called it quits. Former P2P caching provider CacheLogic decided earlier this year to begin using its technology for content delivery instead. (See CacheLogic Pushes Hybrid P2P, CacheLogic Fires Up Its CDN, and CacheLogic Intros P2P CDN.)

When asked why his company switched gears, CacheLogic VP of marketing Gary Croke pointed to the size of the content delivery market, as opposed to its failure to catch fire in the equipment arena. "Content delivery is a $1 billion business in 2007 and will grow to $4 billion in 2010. It's a massive opportunity."

— Ryan Lawler, Reporter, Light Reading

materialgirl 12/5/2012 | 3:05:33 PM
re: Startups See More Cache in Carrier P2P Service Providers labeled peer-to-peer "bad" because they are afraid of it. In fact, peer-to-peer is not "bad", but is the most effective delivery mechanism for significant categories of traffic. Economics therefore dictate that it will become a predominant force in data delivery. Service providers will have to adjust.

In addition to these folks, AKAM bought peer-to-peer technology. Where does it fit in?

What is it about Israel and peer-to-peer? That location comes up constantly in peer-to-peer discussions. I believe the original IM designers were also from Israel.
^Eagle^ 12/5/2012 | 3:05:32 PM
re: Startups See More Cache in Carrier P2P Materialgirl stated:
"Service Providers labeled peer-to-peer "bad" because they are afraid of it. In fact, peer-to-peer is not "bad", but is the most effective delivery mechanism for significant categories of traffic. Economics therefore dictate that it will become a predominant force in data delivery. Service providers will have to adjust."

I tend to agree with her statement that P2P is effective delivery mechanism. I noticed in the article that the carriers are claiming P2P now constitutes "40%" of traffic. What I find odd is that the carriers in some ways are being blind to a great revenue opportunity. Instead of trying to choke off or limit the volume of P2P traffic by using various policy servers and gateways and using DPI; the carriers could choose to enable P2P and even encourage it. I think there is a way to monetize the P2P traffic in a way that would be acceptable to users. Instead of trying to choke of or "manage" P2P, if the carriers simply optimized wavelengths across their edge, metro, regional and LH networks for P2P and sold "portal's" into this optimized set of links.. blind portals mind you that don't care or even look at the P2P traffic. Just ports to a network that ever grows in bandwidth so that if you buy into this optimized subset at a P2P provider you always have "enough" bandwidth. In return for this, carriers can charge a bit more for the "P2P" optimized links or ports than for standard links.. regardless of speed: DSL, T1, all the way through FTTH access up to big ports at OC-x speeds for servers for providers. Instead of fighting the tide, find a way to surf the wave and generate revenue by taking a very small piece of overall transactions either by a transaction based model (bits per month for instance for edge or end users similar to minutes of cell phone.. optimized IP pipes to the edge for gamers and music and video done over P2P).. and / or a "portal" based model that the providers could buy similar to buying a T1 or an OC-3, or even an entire OC-192 or 10GigE lambda on the DWDM transport network. All optimized for P2P to really FLY and enable a whole generation of new users and applications. I can see this kind of enabling strategy could cause even more growth in P2P and the overall revenue streams go up. I also see many applications for business P2P services all the way from banks exchanging key financials transaction data to how content providers supply all kinds of material for distribution to various channels.

So I ask: why not find a way to enable P2P and look at it simply as another pipe from the carrier point of view. Drive massive growth and a new but simple revenue stream for the carriers.

I think lots of gamers, and music users and video users would pay a bit more for their access link if "turbo charged for P2P" was an upgrade option. Like buying premium channels on your cable system. Not just speed upgrades, but links optimized from edge into the "P2P cloud" in the core of the carrier transport network. And if the portals or some equivalent were attractive to the providers of P2P.. and were content neutral.. then I think it would also be a no brainer for many kinds of content providers to simply link into this cloud. Symbiotic expansion of the network by the carriers actually optimizing and enabling P2P and hence growing their own bottom line by simply causing a positive feedback loop. By enabling it.. developers would flock to develop applications and everything would accelerate.

Ride the tide. don't fight it.

sailboat
optodoofus 12/5/2012 | 3:05:32 PM
re: Startups See More Cache in Carrier P2P > So I ask: why not find a way to enable P2P and
> look at it simply as another pipe from the carrier
> point of view.

This is an excellent idea. The carriers aren't even considering this because they do not want to be a pipe vendor (not even a smart pipe vendor). They are blind to the possibilities you describe.

optodoofus
materialgirl 12/5/2012 | 3:05:26 PM
re: Startups See More Cache in Carrier P2P Sailboat:
Music to my ears...

Now we will see just how brave and innovative service provider management can become. If anyone does it, it will be VZ IMHO. They have to change their history and leap into the unknown. T will never come around.

It is monopolist behavior to see 40% of your traffic doing what you don't like at the moment as bad thing and to try and stamp it out. Anyone else would be overjoyed and try to leverage it.
optodoofus 12/5/2012 | 3:05:22 PM
re: Startups See More Cache in Carrier P2P It is a mistake to equate P2P with copyright violation. P2P is just a tool. It has many legitimate uses and applications (e.g., Joost, Skype, etc.) The fact that some people choose to use this technology to steal does not make P2P inherently bad. If we are going to tar all with the same brush, then let's just condemn this whole internet thing since it enables many bad things to occur.

optodoofus
rjmcmahon 12/5/2012 | 3:05:22 PM
re: Startups See More Cache in Carrier P2P It is monopolist behavior to see 40% of your traffic doing what you don't like at the moment as bad thing and to try and stamp it out. Anyone else would be overjoyed and try to leverage it.

Yeah, that's what some New Orlean's police officers did after Katrina. They saw everybody stealing tv's and they got in on the action.

There is no innovation in copyright violations. Nor is there a lasting business model for it. It's behavior that shouldn't be condoned.

Let's try acting like responsible adults instead of pandering to misguided youths. Maybe that's a first step?
longtooth 12/5/2012 | 3:05:20 PM
re: Startups See More Cache in Carrier P2P So we blame the government for making good roads, since it will certainly lead to car theft????

rjmcmahon 12/5/2012 | 3:05:20 PM
re: Startups See More Cache in Carrier P2P So we blame the government for making good roads, since it will certainly lead to car theft????

Think of it this way. Imagine you lived on a farm in a time when cars and trucks were first being made and the local municipality you lived in built some good roads backed by bonds your community secured. Then, looking outside everyday, you saw that 40% of the trucks using those roads were carrying goods stolen from yours and your neighbor's crops and you weren't getting paid anymore. There would be a major backlash against more good roads being constructed anywhere in the country and food production would decline.

Or another perspective. On a recent mountain bike ride my friend and I discovered a tunnel built in the mid 1800s for a narrow guage railroad. It cut through the mountain range making a possible return route for us much easier. Unfortunately, after the attack on Pearl Harbor the locals became so afraid of the Japanese invading their community they blockaded the tunnel by blowing up portions of it. They did this to prevent potential foreigner invaders from accessing their community. 70 years later this tunnel is still blocked and one has to take a long, long way to reach what now has become a very isolated area.

(A side note: Many government regulator types, which includes government affairs personal at major corporations, like to claim that a government can't make any policy stances because of "unintended consequences." The real world is full of unintended consequences so anybody using this argument is making excuses for doing nothing and is acting as a puppet for the status quo.)

So one thought is that if next generations want symmetric, high bandwidth networks which support open access we'll need to have the self discipline not to take what isn't ours by respecting intellectual property rights. If that doesn't happen the network providers may be demanded to enforce the laws in a manner that isolates us from the *free* world.
rjmcmahon 12/5/2012 | 3:05:20 PM
re: Startups See More Cache in Carrier P2P It is a mistake to equate P2P with copyright violation.

I'd agree that P2P not exactly equal to copyright violation.

Can you point me to one in which a majority of the traffic doesn't involve violating intellectual property laws so I can check it out? I haven't been on one in a long while so I'm really not current about what is being carried. In the past when I logged on to a few almost everything being "shared" was unauthorized and strained any interpretation of fair use.

Also, the original poster's suggestion was to make money by "monetizing user behavior" without any other criteria per se. The monetization that lasts would be one that adhered to *legal* user behavior and hence would respect intellectual property laws. (FiOS does that at least.)

Is there a near term solution to the mass piracy being perpetrated by too many who don't experience the guilt required to self regulate this behavior? I don't know of one other than attempting some shame in hopes of influencing ethical behaviors, as futile as that may be. Hence the post, an indictment of behavior by many P2P users and not necessarily that of P2P.

The long term solutions that I perceive are that the media industry works out an accepted DRM system such that market signals can work, the network providers behave as traffic cops, or something in between. I personally believe the closer we are to the former the better, particularly if the DRM system were as open as something like TCP/IP.

Finally, it surely would be nice to hear more people demanding respect for these laws and stopping with the silliness of the RIAA as the devil incarnate, such that the internet, and the freedom that comes with it, represents the traits of mankind we all can admire.

GÇŁEach man takes care that his neighbor shall not cheat him. But a day comes when he begins to care that he does not cheat his neighbor. Then all goes well - he has changed his market-cart into a chariot of the sun. -Ralph Waldo Emerson
t.bogataj 12/5/2012 | 3:05:19 PM
re: Startups See More Cache in Carrier P2P RJ,

While I sure like the way you think ("...we'll need to have the self discipline not to take what isn't ours by respecting..."), I'm afraid we may not live to see it. Too utopian?

T.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE