Hammerhead Retools MPLS Approach

Edge-switch startup Hammerhead Systems Inc. has had a tumultuous year, with layoffs, hirings, and executive changes. The company now has ambitious plans for 2005, with a major shift toward Multiprotocol Label Switching (MPLS) software.

What's in store? Hammerhead now looks to have a new focus in specialized MPLS and virtual private networking (VPN) technology. Specifically, it will seek to drive standards around the technology known as pseudowires -- which allow a wide range of data traffic to be merged over IP networks -- and the recently crafted Dry Martini specification.

Pseudowires are point-to-point connections set up between two networking connection points. Their primary function is to emulate services like Asynchronous Transfer Mode (ATM), Ethernet, Frame Relay, and TDM over a common MPLS and IP network.

Hammerhead admits to having laid off about 20 percent of its staff in recent months, which, combined with some hirings, leaves a headcount of roughly 50. But officials downplay the changes as simply a change in strategy.

"Most of the transition had to do with body composition rather than cuts and adds," says Houman Modarres, director of product management. "Early on, we had a very high investment in hardware. Hardware is still an important part of our plans, but it's a lesser part. We need software, and the software is in new areas for us."

Likewise, change has hit the executive ranks. Departures have included director of field operations Dave Roonie and vice president of sales Cornee Yountz, who was replaced by Jack Reeves in October. On the technical side, CTO John Yu and chief architect Jan Medved are gone as well (see Hammerhead Adds VP Sales and Hammerhead Founder Steps Aside).

Capping it all, the board ousted CEO Joe Sigrist in favor of Peter Savage, who came on board last week. Specific reasons for the change aren't being given, but Hammerhead officials note that Savage has long-standing personal relationships at carriers and at Fujitsu Ltd. (OTC: FJTSY; Tokyo: 6702), which is distributing Hammerhead's products. Savage also knows two of Hammerhead's board members, which certainly helped (see Hammerhead Names New CEO, Headcount: A Pox on Petaluma?, and Hammerhead & Fujitsu Team up).

Hammerhead's HSX 6000 is one of several systems being targeted at the multiservice edge, where MPLS is used to combine multiple types of traffic for transport across a single Internet Protocol (IP) network core. Most competitors are pitching IP routers for this space -- examples being Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Cisco Systems Inc. (Nasdaq: CSCO), Juniper Networks Inc. (Nasdaq: JNPR), Network Equipment Technologies Inc. (net.com) (NYSE: NWK), and Nortel Networks Ltd. (NYSE/Toronto: NT) -- but Ciena Corp. (Nasdaq: CIEN) and Hammerhead are targeting Layer 2 with boxes that would link carriers' older networks into a unified IP/MPLS core.

Hoping to stand out a little, Hammerhead is embarking on a "Layer 2.5" direction driven partly by the hiring of Ping Pan as principal architect (see Hammerhead Orders Dry Martini). Pan is an author of the Internet Engineering Task Force (IETF) draft describing the Dry Martini scheme for pseudowires, and with his guidance, Hammerhead intends to release a flurry of Dry Martini standards proposals during the next year.

The concept of pseudowires, which emulate Layer 2 connections on the connectionless IP network, has been around for a few years. Dry Martini brings it to the Sonet arena, and it's still missing some key pieces, Modarres says. Hammerhead sees that as a chance to seize the reins and become the de facto authority on the technology (see How About a Dry Martini?).

At least five standards are being written, tested, or researched within Hammerhead, all of them related to beefing up Dry Martini with traits found in other protocols. For example, Hammerhead intends to pitch a pseudowire version of Fast Reroute, a feature in MPLS that helps minimize network downtime.

Other RFCs Hammerhead intends to pursue include pseudowire quality of service (QOS); interworking of operations, administration, and management with the IETF's Pseudowire Emulation Edge-to-Edge (PWE3) standards; interworking with Ethernet and MPLS; and a network-to-network interface for connecting across different carriers' networks.

Such a standards blitz is part of a startup's regular playbook. "Sometimes it does work," says Scott Clavenna, chief analyst with Heavy Reading, Light Reading's paid research arm. "If you look at other pseudowire standards, a number were driven by Axerra Networks and RAD Data Communications and have been adopted -- mostly the ones around circuit emulation."

One major question will be whether the IETF is receptive to Hammerhead's ideas.

"The thing with Dry Martini is that it lets you create 'tunnels' in Sonet and other 'sub-IP' networks to provision packet traffic across. But these aren't really LSPs [label switched paths] in the classical sense of the word, so the IETF doesn't quite know what to do with it," Clavenna says. He adds that, should the IETF turn its nose up at some of the proposals, Hammerhead might find a better audience at a group such as the MPLS/Frame Relay Alliance

Separately, Hammerhead noted in a release today that the HSX 6000 is in lab trials. Rumor has it MCI Inc. (Nasdaq: MCIP) and Savvis Communications Corp. (Nasdaq: SVVS) are both having a look, but officials wouldn't comment on whether that's true.

— Craig Matsumoto, Senior Editor, Light Reading

For further education, visit the archives of related Light Reading Webinars:

firstmile 12/5/2012 | 12:59:28 AM
re: Hammerhead Retools MPLS Approach I don't think that Savage brings strong relationships with anyone other than the board member that placed him in the position. When is the last time that he was involved with selling something to a carrier? Applied Digital Access was a LONG time ago...
Ringed? 12/5/2012 | 12:59:26 AM
re: Hammerhead Retools MPLS Approach Firstmile,

Relationships are what they are.. You have one or you don't. In today's climate arbitrary BoD placement for the sake of placement doesn't work.

Even with a personal realationship; there is not a VC that will take a chance on a weak push...

A Board member should never sell directly(said product) to the executive staff of any (potential) company. This type of sale would need to be behind the scenes; if at all and should only constitue the 1st (business transactions) a startup should see.

The merits of a sale in a start up requires the tenacity of a person that has: One, dealt with the customer before, thus comfortable. Two, a customer that wills to take a chance, (lately due to price). Three, Looks for a vendor that will serve the customer the way they were served by LU, NT & CSCO pre-bubble.

In short. The executive staff can ruin this company more quickly than the Engineering Staff, Marketing Staff, or sales staff ever will.

Can we all get on a Ringed?
ChatEnChapeau 12/5/2012 | 12:59:25 AM
re: Hammerhead Retools MPLS Approach Factually incorrect.

Drafts are "individual submissions" until the working group they are submitted to accept them. That's when they become somewhat real - there is a rough consensus of people in the working group that think there is merit in the draft.

Dry Martini isn't there, and isn't going to be. We suffer when we have multiple ways of doing things, and unless there is some compelling reason to do so, no-one is going to rewrite their original draft-martini implementation to accommodate some parvenu.

Truelight1 12/5/2012 | 12:59:25 AM
re: Hammerhead Retools MPLS Approach Pelease,

Savage does not have a clue about this market or products. The BOD is done as is HammerDead.

standards-watcher 12/5/2012 | 12:59:22 AM
re: Hammerhead Retools MPLS Approach Ping Pan brought up his "Dry Martini" approach at an IETF meeting, but it was immediately shot down.

The problem was not the IETF not being receptive, it was just that the idea was too trivial.
pingpan 12/5/2012 | 12:59:20 AM
re: Hammerhead Retools MPLS Approach Ping Pan brought up his "Dry Martini" approach at an IETF meeting, but it was immediately shot down.

The problem was not the IETF not being receptive, it was just that the idea was too trivial.


"The proof of pudding is in the eating". My job is to make carriers doing better, by providing running code and working product. After watching the fate of the original 2547 and MPLS-ping, I have no interest to please any standard body. If anything to get shutdown, rather let my customers to tell me so.

BTW, sorry about making it too simple. It was not meant for putting on routers. It was meant to add on simple access device, transport gears and other non-IP centric devices.
firstmile 12/5/2012 | 12:59:18 AM
re: Hammerhead Retools MPLS Approach Ringed,

What world do you live in? "Even with a personal relationship, there is not a VC that will take a chance on a weak push."

Take a look around the marketplace.

Do you think that Santa is coming down the chimney too?

ATMRules 12/5/2012 | 12:59:10 AM
re: Hammerhead Retools MPLS Approach Company will be 3 yrs old next month, gone thru all the typical, restructuring, VP's gone, new blood added, down to 50 people and now focusing on new market.."pseudo wires"....ok whatever you say....this a classic telecom startup with a good (original) idea but a dollar short and a day late to a tough market...What RBOC is going to put your product in a live network? I think your TTL is 6-12 months at best depending on the size of your bank account and how much patience your VC's have...or you could get lucky and some other vendor will purchase you........Good Luck!!!

ATMRules 12/5/2012 | 12:59:01 AM
re: Hammerhead Retools MPLS Approach One more thing, how about re-tooling your website? It's rather lame don't you think?

ATMRules... IP will never Have the Qos of ATM, close but no cigar...(yet)
firstmile 12/5/2012 | 12:59:01 AM
re: Hammerhead Retools MPLS Approach Classic...can't agree more...
"But you never know, maybe someone will buy us for $400M within the next 8 months...I hear that Mr. Savage has great relationships with the carriers."


Sign In