Ethernet equipment

Cisco Feels Switching Pains

Even as Cisco Systems Inc. (Nasdaq: CSCO) prepares to possibly cut more businesses (Linksys and WebEx have been mentioned recently), it's facing some issues with one of the units it has to keep.

The question is how quickly the Ethernet switching business, as much a Cisco franchise as French fries are to McDonald's, can return to growth, and whether it can retain the same margins as before. On Cisco's recent earnings call, CEO John Chambers mentioned switching -- where revenues are down 9 percent from a year ago -- as one of the company's biggest challenges.

Competition is a primary reason. "Switching is an area Cisco has less control over, given the rising competitive forces," wrote analyst Ittai Kidron of Oppenheimer & Co. Inc. , in a recent research note.

At the high end, switch chips from Broadcom Corp. (Nasdaq: BRCM), Fulcrum Microsystems Inc. and Marvell Technology Group Ltd. (Nasdaq: MRVL) are giving rise to a tough new generation of 10Gbit/s switch boxes. Consultant Nick Lippis of Lippis Enterprises counted nine top-of-rack switches being announced at Interop earlier this month.

And in plain enterprise switching, competitors -- HP Inc. (NYSE: HPQ) gets cited a lot here -- are helping drive down prices, which in turn is helping to lower Cisco's prized margins. The company's overall gross margin has fallen to 63.9 percent from 70 percent a couple of years ago, one of the factors that's led investors to push Chambers for changes.

What does it all add up to? Possibly, a chance for a competitor to put a long-term dent in Cisco's market share, much the way Juniper Networks Inc. (NYSE: JNPR) did on the service-provider side years ago. "A combination of factors has opened up opportunity for a second vendor to enter the discussion," says Scott Dennehy, an analyst with Technology Business Research Inc. (TBR) .

It's possible, then, that Cisco might have to accept lower switching margins as a fact of life. That's not the same thing as becoming a commodity switch vendor, Dennehy tells Light Reading: "I don't think they'll ever accept margins as low as HP would accept." But he suspects the pressure on Cisco is here to stay.

He thinks Cisco's best response might be to concentrate on the high end, not only because margins are better, but because price isn't always the trump card in those markets. Education, for example, might be an area for Cisco to emphasize less. "If you go into a school and say you can save $1 million with HP, they're going to go with that," Dennehy says.

Another reason to focus on expensive switches is that Cisco's big architectures -- grand schemes that include interlocking groups of products -- might not appeal to some switch customers. With role models like Amazon Web Services Inc. and Google (Nasdaq: GOOG), some cloud and data center owners might wonder if the same model -- building commodity networks out of throwaway boxes -- could work for them as well.

Follow the headlines below for a quick summary of Cisco's latest reorganization so far.

— Craig Matsumoto, West Coast Editor, Light Reading

Pete Baldwin 12/5/2012 | 5:04:57 PM
re: Cisco Feels Switching Pains

I wouldn't expect Cisco to just walk away from the low-end switching market, but it does make sense that they should eventually concede defeat (or disinterest) there, doesn't it? I'm thinking not only of the heavy competition and the easy availability of good Ethernet switch chips, but also of the possible effects of OpenFlow.

5+ years ago, Cisco was talking as if they saw this day eventually coming. But their answer was the architecture thing, and I wonder if enterprises will gravitate towards something more like a disposable switch.

crazy4geek 12/5/2012 | 5:04:56 PM
re: Cisco Feels Switching Pains

The pricing pressures of low to mid end switching is going the way of the consumer market.  In these markets, particularly public sector, pricing is king and they are less likely to buy into the Cisco architecture play.   Combine this with a large reduction in public sector spending and it may be time for Cisco to focus on their high end switching line.  I would think Cisco would remain in the low to middle end switching, but with less emphasis.  Simultaneously, Cisco can focus on all the things mentioned regarding HP competitiveness and cut the inevitable drop in margins off at the pass.  H3C had the Asia market cornered prior to the HP purchase.  Will HP take over high end switching faster than Cisco can find ways to be more price competitive?  We all have a front row seat to see how that plays out- my bet is on Cisco.  

Marc_A 12/5/2012 | 5:04:56 PM
re: Cisco Feels Switching Pains

I worked for Cisco for a decade and a half, and I'm now competing against Cisco at HP.  The challenge that my former employer faces from HP is more than HP's (H3C's) use of merchant silicon vs. Cisco's custom ASICs.  I’ll offer you some historical background and then list the differences that are resonating with customers.  I realize that the latter will sound like a sales pitch, but I am honestly just trying to describe the greater challenge that Cisco is facing.

Though referred to as 3Com, HP acquired H3C in April 2010.  H3C was the joint venture between 3Com and Huawei between 2003 and 2006, when 3Com bought Huawei’s share and became the sole owner of H3C until HP’s acquisition.

H3C had a clean sheet of paper when they developed their enterprise switch line.  In addition to their new hardware architecture, they re-wrote the Comware O/S and developed their multi-vendor management system, IMC.

Over time, H3C surpassed Cisco as the number one market share holder in China, which led to the “China out” policy, to sell H3C products globally.  HP then evaluated all network vendors (except Cisco) for acquisition and chose 3Com (H3C), which became HP Networking’s (HPN’s) A Series products.

The use of merchant silicon is indeed a key factor from a margins perspective, and it also results in greater energy efficiency when compared to Cisco’s ASICs.  There is more that resonates with prospective customers though, which I'll describe through the following points.

HPN’s A Series products have a common hardware architecture, with a single open standards based O/S (not just a common CLI like IOS).  Comware is also roughly 90% similar to IOS in its CLI and command structure, so CCIEs and CCNAs have little difficulty mastering it.

Unlike Cisco, HPN does not use a feature license model.  All products are fully feature enabled out of the box, which includes hardware based dual stack IPv4 / IPv6 for all products.

HPN’s edge switches have a lifetime warranty, versus Cisco’s new limited lifetime warranty.

Both Cisco’s Smartnet and HPN’s Carepacks are priced at a percentage of the products’ prices.  Because HPN’s products are 30% to 50% less expensive than Cisco’s, the on-going service and support costs are also significantly less, which becomes very apparent when a customer considers their 3 to 5 year TCO.

I was working for Cisco in 2001 during their first RIF, and the company is very good at putting the brakes on and controlling their costs.  The landscape was different back then though, because Cisco was the vendor who survived the switch wars of the late ‘90s.

To prevail as well this time, Cisco would have to switch to merchant silicon (as they are with their project jawbreaker), move to a single O/S, jettison their profitable feature license model, and slash their services pricing.  That would be a very different company than the one that lines Tasman Drive today.

ycurrent 12/5/2012 | 5:04:46 PM
re: Cisco Feels Switching Pains

As always its a trade-off between market share and margins, and especially in switching, it makes sense to compete and preserve higher margins business than to protect a lower margin business.  Absent from the comments is looking to where the growth is/will be.  Often these posts are very US-centric. If there is much growth elsewhere (as I suspect there is), then Cisco should still find healthy margins (and competition) in enterprise markets outside of the US, without focusing on verticals like public/education, where growth is slow/negative? and prices are pressured downwards.

Pete Baldwin 12/5/2012 | 5:04:43 PM
re: Cisco Feels Switching Pains

Marc, thanks for taking the time on that posting. Great analysis.

Sounds like Cisco can be needled from a lot of different directions at once. The single-OS model certainly is one (and it's a model that just makes so much sense - maybe it just appeals to the control freak in me).  I've never learned much about the sales side -- the warranties and feature pricing -- so that's interesting to hear as well. Thanks.

Sign In