Cisco to Pay $89M for Procket Assets

As anticipated, Cisco Systems Inc. today announced it will buy the key engineering staff and entire intellectual property portfolio of core router vendor Procket Networks Inc. in a cash deal worth $89 million (see Cisco Acquires Parts of Procket and Cisco/Procket Deal Imminent).

Cisco isn't providing any further detail at this point, but a spokesman notes that it "isn't going to buy the company lock, stock, and barrel. We won't acquire any products. We'll buy the IP and Procket's talented engineering team."

But Cisco isn't taking on all of Procket's engineers. The spokesman says there's no breakdown of who is or isn't joining Cisco's Routing Technology Group, which is managed by senior vice presidents Prem Jain and Mike Volpi.

There's also no further detail on what the "select assets" referred to in the news release are, though again the spokesman presses the point that these will not include any Procket products. He adds that Cisco will not assume any of Procket's liabilities.

In its official release, Cisco notes that the "purchase will add a rich intellectual property portfolio and a team of proven silicon and software architects to Cisco's industry-leading routing technology and products."

Volpi is quoted as saying: "The addition of Procket's engineering team to Cisco offers a unique opportunity to accelerate development of silicon and software across Cisco's next generation routing portfolio. Procket has some of the world's foremost designers of sophisticated silicon, software, and network systems with an average of over 15 years of experience in their respective industries. Adding this talent to Cisco's world-class engineering team will help drive continued innovation in network infrastructure, and that is good news for our customers and the industry at large."

Cisco recently unveiled its own core router after years of stop-start development (see Cisco Unveils the HFR). It's no doubt worth $89 million to Cisco simply to take out a potential competitor.

The deal is subject to standard acquisition conditions and regulatory approval, and is expected to close in the first quarter, ending in October, of Cisco's 2005 financial year.

In pre-market trading this morning, Cisco share price was down by 30 cents, or 1.3 percent, to $23.58.

— Ray Le Maistre, International Editor, Boardwatch

<<   <   Page 27 / 27
echo2 12/5/2012 | 1:30:25 AM
re: Cisco to Pay $89M for Procket Assets Perhaps they also had it running on real hardware for more than a year.

All the major carriers had this in their lab for a year and no one bought and deployed? I think you can understand this would be a warning sign for a potential buyer. And for some, one year of lab testing is still not quite enough. It's sad, maybe unfair, but some competitors just happened to have a more field proven system.

The fact that you didn't get picked doesn't necessairly mean the suitors don't understand the value of a core router. You certainly had the sexiest system, but maybe it just wasn't the one most mature and carrier-grade one which is what some customers need.


Perhaps it was lack of scalability and lack of high availability.

A sexy product or a product that can beat the competition by a few percent in the benchmarks just ain&#8217;t going to cut it in the core router market. Think about it. You are responsible for a large carriers national IP backbone. You would be a total dumb ass, and on unemployment, if you jeopardized a national backbone because you thought a particular core router was sexy. You can get away with that at the edge, where if the vendor disappears you can just walk over to your switch closet and throw out the old router and install the new one in a few minutes.
st0 12/5/2012 | 1:30:24 AM
re: Cisco to Pay $89M for Procket Assets "The fact that you didn't get picked doesn't necessairly mean the suitors don't understand the value of a core router. You certainly had the sexiest system, but maybe it just wasn't the one most mature and carrier-grade one which is what some customers need."
or otherwise,it just "waited until just prior to fire sale".... understand that support for testing require a lot of resources, and you can not get out of it easy (without lost "face")... once your gear in the test site, the "customer" got upper hand... time is not on the start up site.... "test until they drop" is good way to get something cheap just prior to the "close door sale"...you have to go through the bidding war at "close door sale"... Cisco may not win the game (just in case Huawei wanted it..hahaha...

jimbostevens 12/5/2012 | 1:30:22 AM
re: Cisco to Pay $89M for Procket Assets I was reading an article stating that Cisco will spend about $685,000 per Procket engineer.

Does anybody know the length of time specified in the noncompete clause the engineers are comitted to and which of Procket's top engineers are NOT going to Cisco?

Also, does anybody have an idea how Juniper's next generation router will compare with the HFR/CRS? Other than price, or maybe business relationship, why would a carrier buy an HFR from Cisco rather than a proven product like the T-640 from Juniper that has a minimum of 60 customers?
signmeup 12/5/2012 | 1:30:22 AM
re: Cisco to Pay $89M for Procket Assets jimbo (or more likely Juniper HR/PM guy),

I'm sure you are not going to get that information from this board. Why, other than trying to hire them into Juniper yourself, would you want to know specifics about how long the non-compete is for, and which "top" engineers would not be going over???

Is juniper that scared they have to resort to trolling on LR for talent? Man, how have times changed...

As far as the $$$/engineer, you are not counting the cost of the IP associated with the deal. Even though the real deal was for the people, keeping the IP out of competitors hands was just an added bonus. BTW, how DO you develop a shared memory architecture that can support 40Gbps per slot cost effectively? I dunno, ask Cisco as they now own that IP... How to create a 40G programmable network processor without it using [email protected],124 BTU/Hr like the HFR? Once again, ask Cisco, as they now own the engineers at Procket that did it...

Sounds like you might be a little worried that your next generation product may not be all that next generation when compared to the Cisco/Procket combination.

jimbostevens 12/5/2012 | 1:30:21 AM
re: Cisco to Pay $89M for Procket Assets Signmeup, you're way off base! I own Juniper shares and that's my only connection to the company. I'm sure top router people know each other and know how to reach one another without posting here. My questions are from the standpoint of an interested observer and investor and nothing more. Your assumptions are foolish!

Having said that, I do think that Procket had a great team of people and technology. Unfortunately, for numerous reasons, including mismanagement, poor market climate, infighting etc., they failed.

Cisco purchased Procket at a fire-sale price in my opinion. However, they screwed up by not purchasing Procket's technology and engineers much sooner. Potentially, Alcatel and Huwaei were the biggest losers in this deal.

You sound like a Cisco employee worried about Procket's staff jumping ship. Cisco needed to purchase Procket to avoid falling even further behind Juniper in Technology. Combining Cisco and Procket's technology will be a complicated process. It remains to be seen how well the two teams blend together.

My bet is that Juniper will continue to gain market share during the next 6 months and an investor will make more money owning Juniper rather than Cisco. Juniper's one of the most interesting startup stories around and it'll be fascincating to see how things develop looking back 5-10 years from now...

P.S. Tony Li would have been hired by Juniper by now if I had any influence. Good luck Tony!
gotman 12/5/2012 | 1:30:18 AM
re: Cisco to Pay $89M for Procket Assets jimbo,

BS you don't work at Juniper. You just don't have any influence.. Juniper will be running scared from T-640 to J-series!! Watch the gap.
Did you not read the T640 vs 12xxx thread? Good luck.
Truelight1 12/5/2012 | 1:30:18 AM
re: Cisco to Pay $89M for Procket Assets Your sucking for Tony would have resulted in selling Juiper stock
turing 12/5/2012 | 1:30:13 AM
re: Cisco to Pay $89M for Procket Assets >All the major carriers had this in their lab for >a year and no one bought and deployed? I think >you can understand this would be a warning sign >for a potential buyer. And for some, one year of >lab testing is still not quite enough. It's sad, >maybe unfair, but some competitors just happened >to have a more field proven system.

You guys are totally missing it. You're still arguing best router or most field proven blah blah. Technology doesn't sell routers. Sales sells routers. Sales through relationships, dominance, or pricing. Technology just helps you get looked at (in the lab). But the lab guys don't make the buying decision - top management does. Buying core routers is a strategic decision, not like buying a cell phone. Unless it is fundamentally a new type of product which they can't get elsewhere, they will go with trusted (legacy) names. The days of gambling are over.

Juniper succeeded because they had great timing - a time when technology could trump comfort, and even then they only got the carriers that were already disgruntled at Cisco. Now they are one of the comfort vendors. Procket just didn't have enough differentiation to overcome these - most carriers don't feel they need a new core (yet) and the procket box did not create some funamental change to making revenue. It was not the fault of procket engineering, management, or technology.
routingfool 12/5/2012 | 1:29:37 AM
re: Cisco to Pay $89M for Procket Assets Since there has been no adequate response from Juniper on the miercom report, I can only conclude that there tests are valid. It would be nice to see the test configs
spaceman_spiff 12/5/2012 | 1:27:45 AM
re: Cisco to Pay $89M for Procket Assets Juniper might have the same problem, that they don't know the configs. It was said a few times already, that it is always possible to find some kind of weird configuration for every box, which causes strange behavior.

All the reordering was already discussed, so it is a non-issue. Also the HOLB issues discussed in other threads might only happen in case of massive overutilization of interfaces (test with up to 175% in our case never brought down the wirespeed performance of a neighboring interface on the same PFE). So in this case you have other problems.

This leaves us with this access list. I have a potential explanation. Juniper offers the option to count packets based on various selection criteria. One counter uses one area of memory, which is pretty obvious. It seems to be possible by heavily accessing this single counter from many interfaces that the memory interface gets overloaded and slows down the forwarding. So they send traffic, which always matches the access list criteria, from a tester to who knows how many interfaces. This results in updating this single counter. For counting all the traffic coming in through interfaces, each interface has a counter, so there is no purpose in doing things that way.

I've done consulting for more than five years now (also dealing with Juniper) and this configuration is far off anything anyone would ever use. Cisco spreaded an internal report to a few carriers last year, which used the same type of configuration. The people I talked to didn't take it serious by that time, because of the obvious irrelevance for real networks.

From my point of view this seems to be a pure marketing activity initiated by Cisco. Maybe their CRS-1 hasn't got enough attention because of an established product. Now Cisco is hit with what they did in the past. They had the GSR and it still has incredible sales volume, because it is established and gives competitors a hard time to get into a network. Juniper established their T640 and now Cisco has to try to re-enter the market with their new product. Now they use their old cooked up tests and put it to a wider audience in order to damage the image of the established T640. Juniper did similar things to Cisco in the past and they are doing it today (I won't comment on details!)

No box is without issues. No marketing campaign is without issues.

I don't think this Miercom report will change anything. Depending on who you talk to, you will get one of two answers:
1. Talking to the Cisco guy, he will reply: The T640 has some major design flaws and this will cause you problems if you buy it.
2. Talking to the Juniper guy, he will reply: Cisco hasn't done itself a favor, because these tests are not relevant. Because they haven't found something real, they try to make something up. They don't have a superior product and they are getting nervous, because of the success of the T640.

Would you buy a C128xx now? Or a T640? Or a CRS-1?
<<   <   Page 27 / 27
Sign In