Sockeye Spawns Service

The IP routing optimization business is heating up. Sockeye Networks, which raised $28 million in a single round of funding earlier this year, today launched its GlobalRoute service and announced its first customer.

Focal Communications Corp. (Nasdaq: FCOM), a national communications provider, has successfully concluded its trial of GlobalRoute and has deployed the service in its IP network. Sockeye says that eight other companies are either currently testing the service or are about to fully deploy it.

This announcement comes just as the IP routing optimization market is getting off the ground. Sockeye’s two main competitors, NetVmg Inc. and RouteScience Technologies Inc., have both recently announced products of their own (see NetVmg: A Bandwidth Cost Cop and RouteScience Soups Up IP ).

But unlike RouteScience and NetVmg, Sockeye is using a service model instead of selling systems outright. The service will collect data on Internet performance using technology licensed from Akamai Technologies Inc. (Nasdaq: AKAM). Sockeye is the only company with rights to Akamai’s network of more than 13,000 servers on over 1,000 networks. The exclusivity contract between Akamai and Sockeye is good for five years, says Jay Seaton, chief marketing officer for Sockeye.

”The fact that Sockeye has access to Akamai is key,” says Elisabeth Rainge, director of network management for IDC. “It brings more dynamic knowledge into the picture, and that’s important. Akamai is an important starting point.”

Using information from Akamai, Sockeye identifies congestion at particular points of presence or peering points and then uses its own proprietary algorithm to collect more specific data regarding that segment of the network. Based on this information, GlobalRoute can do a couple of things. First, it can generate custom reports for enterprises or service providers, which can be used for future network planning. It can also be pre-programmed to automatically reroute traffic to a less congested link if a threshold has been exceeded on a particular route.

Large enterprises and service providers already use collected data to optimize their routes, but using the Sockeye service or an appliance from NetVmg or RouteScience automates that process, ultimately making it more cost effective, says Rainge.

Sockeye claims that the GlobalRoute service improves 70 percent of BGP (Border Gateway Protocol) routes and lowers bandwidth and network management costs by up to 40 percent for businesses and service providers that use more than one connection to the Internet.

While Sockeye’s relationship with Akamai may be its biggest competitive differentiator, the relationship has raised some eyebrows. According to a Reuters report published in August of this year, Sockeye, originally called Arriva, was actually formed as a subsidiary of Akamai. In January 2001, Akamai sold a majority stake in Sockeye to outside investors for $28 million cash. The sale included a five-year licensing deal that requires Sockeye to pay Akamai a minimum monthly license fee of $1 million plus royalties based on Sockeye's revenue from using Akamai's technology, says the Reuters report.

Sockeye accounted for 11 percent of Akamai's $83.3 million in revenue during the first half of this year. Akamai holds a 40 percent stake in Sockeye; its chairman George Conrades sits on Sockeye's board of directors; and some of Akamai's largest investors -- Baker Capital Corp., Battery Ventures, and Polaris Venture Capital -- are financing Sockeye and continue to hold large stakes in Akamai.

Some sources have charged that Sockeye was formed to reinforce Akamai’s business, which has still not turned a profit. But a spokesperson for Akamai denies these claims and says that Sockeye is a completely independent company that simply licenses its technology.

Regardless, GlobalRoute is not a cheap service, and Sockeye will have to prove to customers that it is really worth the cost. Pricing is scaled based on the number of locations and the number of links at each location, with higher rates for higher-capacity links. For example, two fractional T3 (45 Mbit/s) lines in a single location costs $4,500. Two OC3 (155 Mbit/s) connections in a single location would be about $12,000. There is also a $6,000 non-recurring charge per location. Discounts are offered for multiple locations and multiple links.

Jay Seaton, chief marketing officer, says that some high-end users with high traffic rates would likely require at least three locations each with three links per location. For these customers -- large enterprises with multiple locations and multiple providers or large service providers -- the monthly cost could soar from $20,250 for three fractional T3 links to as high as $54,000 for three OC3 links at each location. Also factor in the one-time fee of $18,000 -- $6,000 for each location where gear is installed. For three locations with three OC3 links per location, the price for the year could be as high as $666,000.

The bottom line is that optimizing IP routes isn’t cheap. Even Sockeye’s competitors are expensive. RouteScience PathControl ranges in price from $140,000 to $250,000, depending on configuration, and that is per device. NetVmg sells its Flow Control Platform at a price starting at $20,000 but ranging up to hundreds of thousands of dollars, depending on the size of the installation.

— Marguerite Reardon, Senior Editor, Light Reading
WeNoNeedNoStinkinMPLS 12/4/2012 | 7:29:49 PM
re: Sockeye Spawns Service The entire notion of this supposed GǣnewGǥ market place is nothing short of a joke. From Route Science, Sockeye, NetVMG, and Opnix (formerly known as the "Tier 0 backbone") now being realigned into this space are all in a non-existent market. This is the classic example of the tail wagging the dog with many people involved that donGt understand the basics of BGP and the fundamental way in which BGP policy can be used/influenced and the limitations.

First, BGP is not symmetrical with how traffic traverses across domains. The way traffic goes toward the destination is NOT guaranteed to be the same path on the return back toward the source. When a given enterprise or service provider is multi-homed. More often than not itGs ASYMETRICAL. Anyone who has any experience with BGP knows that you may try and take the path of least congestion going outbound, but the return path is based on the other autonomous systems and the remote ASGs BGP policy that the traffic traverses through.

No matter how each of these companies try and "market" their story and its quite obvious that what you have is VCs that clearly have very little experience in this space and were sold a bag of goods. What you see is teams of VCs trying to make sure they have Gǣa playGǥ in this Gǣnew marketGǥ.

Next time you see one of these companies on a show floor or in a private demo, ask them to see their 10 line PERL scripts. If you care learn some basic PERL scripting, I can offer classes for quite a bit less than the cost of this ridiculous idea.
skipjacks 12/4/2012 | 7:29:48 PM
re: Sockeye Spawns Service Looks like a fantastic opportunity to create the modern version of broadcast storms, routing loops, throughout the network. Anyone who is a service provider or makes equipment with BGP should watch out!
green 12/4/2012 | 7:29:45 PM
re: Sockeye Spawns Service 1: the parent (Akamai) is yet to prove its business model. Building it off Akamai is shaky at best
2: If there is indeed a business case I would rather go with the competitors who have hardware based solution.
buliwyf 12/4/2012 | 7:29:43 PM
re: Sockeye Spawns Service Sockeye began as little more then a way for Akamai to manufacture new revenues by funding a spinout that becomes a customer of akamai.
Not only licensing fees but also another 3.7M for technical development.

Same company that invests $5M in a colo facility of one of its senior officers (net.eng perhaps?) and also purchases 600k worth of services from this company each half. (from 10-Q)

The amusing thing is that companies like sockeye appear to be aiming at corporates. Given that knowing which upstream peer you hand off to provides minimal if any performance gains (perhaps if the entire internet was running the service one could argue that there may be performance gains), why is a corporate going to buy this service?

And how is focal's cash position anyway?
Fiberfreak 12/4/2012 | 7:29:37 PM
re: Sockeye Spawns Service ".....what you have is VCs that clearly have very little experience in this space and were sold a bag of goods...."

Right emotion but I think VC's are not *sold* bad ideas, but actually drive this process.

Some VC's will do ANYTHING to try and create some hype around a purported "new" market - remember that only the early players in a pyramid scheme make any money by getting out when the hype peaks. They take advantage of the VC herd mentality - remember the DOT COM'S; VC's are so afraid they will miss out on yet another way to flip a few bucks, so the herd starts moving and the early players make a buck and split.

Why don't VC's get some real expertise in house and create some worthwhile long term plays instead of taking a fundamentally flawed idea and trying to flip it without regard to the longevity (and future lost value) of another "new" market gone bad? Haven't there been enought screwed employees and investors out there already? Or maybe thats what it's all about - quick bucks at someone elses expense. After all, real expertise costs money.
Sign In