VMware, Cisco, Nicira: The Devil You Know

Disney-Chicken-Little-Sky-FallingAs every one of you knows by now, VMware announced its intentions to acquire Nicira for $1.26B last Monday.  I’ve waited a few days to post this, just because I’ve been fascinated by the conversations that have happened in response.  To recap, here are some of the conversations, and a description of the general talk track:

twitter-lol-600Twitter was equally bombastic, and quite a bit more humorous.  Here’s some of the ones that I can share, but trust me, my DM list looks like open mic night at the Improv…

  • Joshua McKenty @jmckenty: For only $1.26B, VMware joins #OpenStack (by buying Nicira, and the PTL of Quantum). A gold sponsorship would have been 0.0000397 of that. (one of my personal favorites!)
  • Brad Casemore @BradCasemore: Now that VMware has acquired Nicira, the EMC-Cisco divorce could be imminent.
  • Rohit Sharma@rohit_x_: Nicira acquisition a sign of VMware’s vulnerability and weakness, not strength of Nicira’s technology/products.
  • Florian Otel‏ @FlorianOtel: That odd sound you hear ? It’s N1kV crying his eyes out #VMware #Nicira
  • Alessandro Perilli@a_perilli: DynamicOps first, now Nicira. VMware is preparing to go big on multi-hypervisor management. The time has come.
  • Joe Onisick@jonisick: Yep #VMware intends to acquire #Nicira. Now they’ll be competing full on with #Cisco
  • Damon McKay@DMcKay: VMware acquisition of Nicira.. that awkward moment when the wife (Cisco) finds the babysitter sitting at the breakfast table. What next VCE?

This acquisition and some of the conversations are naturally of interest to me, since I work for a company owned by Cisco, VMware and EMC.  Any time new technology is brought into the family, or new tensions, it’s something that will affect us individually and collectively.  That said, I’m just a worker-bee, and I don’t have any insight into the machinations that went on behind the scenes.  Let me say that again: I don’t have any more insight into this acquisition and it’s impact than anyone else who works in the industry.

That being said, I don’t get it. I understand what Nicira does.  I understand the threat that such a company could provide to Cisco, especially in a situation where the hardware switch business that Cisco counts on could be impacted.  I understand the efforts Cisco is making around programmability (SDN, in Cisco-speak) and how much of that was in response to advances made by companies like Nicira, Big Switch and Embrane.  I understand the existing overlap in the VMware and Cisco product sets, idontgetitand the tension created when VMware went deep into the virtualized switching and security spaces (DvS vs. N1000v, vShield Edge, vShield App vs. VSG), and then tightly integrated those products into vCloud Director.  I understand how important the Nexus 1000V is to the architecture that Cisco envisions as they virtualize more of their security and content management products, and how VMware’s advancing of features in their included Distributed vSwitch threatens that virtualization process.  I understand that Cisco attempted to buy Nicira and was rebuffed, and having VMware end up with the prize is a blow to Cisco’s ego and contributes to the perception that Cisco isn’t a place where start-ups want to land.  I understand all that.

What I don’t get, is how it’s a strategic problem for Cisco.  On the contrary, I would think that Cisco would be ecstatic that the technology ended up with VMware.  Why?  Here’s my two primary reasons:

  1. VMware, especially in the short term, has a very different need from the Nicira product than Cisco was originally worried about.  VMware is not going to be an x86-based switch provider, they are going to continue to be a hypervisor provider and virtualized infrastructure (software defined data center?) management company.
  2. VMware is not going to recognize revenue tomorrow that Cisco is recognizing today.  The gradual erosion of the top-of-rack switch market is a process that has been going on for years.  This acquisition isn’t going to change that, or accelerate it.  If anything it’s going to slow it down, because it pulls a prominent and open-source friendly SDN player off the market at a time when growth is rapid.

vmware-nicira122So what could VMware do with Nicira?  I bet they could use it to help with the extensibility and functionality of the existing vSphere networking stack.  I bet they could use it to help build some of the multi-site networking features required for stretched and metro clusters directly into the hypervisor, driving down the cost and complexity of those designs just as they are starting to become more popular.  I bet they could use it to create a virtualized network control plane to better help manage traffic paths between different elements within the stack.

From Cisco’s point of view, aren’t there a dozen WORSE places for Nicira to have ended up?  Wouldn’t it have been worse if a direct competitor like Juniper or Palo Alto or Fortinet or someone who actually competes at a 1:1 product level with Cisco had been involved?  Wouldn’t it have been worse if IBM or Dell were able to start driving switching into generic appliances?  I feel like it would.  I feel like Cisco would be bunkered down trying to figure out when the sky was going to fall, and the market would have started selling off in droves.

Better-The-Devil-You-KnowVMware and Cisco compete directly in many places even outside of the pure network space, particularly around the orchestration and service catalog piece.  This isn’t something that’s unique to the VMware/Cisco relationship, it happens all the time between large enterprises with strategic alliances.  Hell, Cisco and EMC own my company, and the Cisco/NetApp relationship is still alive and kicking, and EMC/Brocade is as strong as ever.  This is how companies work.  I hate the word “coopetition”, but there you go.

It does seem like there’s a LOT of invested effort put into playing up any and all difficulties between VMware, EMC and Cisco.  I guess that’s what happens when you sit at the top of the hill; everyone takes pot-shots.  What I know from personal experience is that our Chairman and CEO came from Cisco, our president and VP of Sales came from EMC and the level of talent all up and down the org chart that came from the parent companies is impressive.  I know that every formal statement made by anyone involved, be it John Chambers, Joe Tucci, Paul Maritz, Rob Lloyd or others has been overwhelmingly positive.  I know that the numbers every quarter are orders of magnitude improvements YoY.  I know the percentage of global Cisco data center revenue that VCE contributed.  The shadows that are being jumped at don’t exist, in my experience.

In its own special place of honor down here, I wanted to call out possibly the most poorly referenced and asinine of all the articles written in response to the Nicira deal, Barb Darrow’s tour-de-force of ad revenue-generating slapstick, “Vblocks – here today, where tomorrow?” over on GigaOm.com. Whether you like the conclusions drawn by the posts I referenced above, you can (usually, kinda) see the logic that was used to arrive there. Not so with this one. With a headline that would impress even the Register, Barb dives into a litany of unsupported conjecture and unconnected quotes. Let’s look at my favorites:

Now that VMware has bought Nicira and its software-defined networking expertise, doubts about the future of the VCE Vblock effort have multiplied.

No where in the article are those doubts listed, or is proof of their multiplication shown, but it does make for a good opening line, doesn’t it?

What’s new is that with Nicira in the fold, VMware gets more directly competitive with Cisco, further stressing the partnership.

I’d love to see the reasoning behind this statement. As I pointed out above, there are dozens of people who COULD have bought Nicira that would have been a much, much more direct threat to Cisco.  Even if Cisco does look at VMware differently now, how does that affect the partnership?  Cisco partners with EMC and NetApp directly on competitive offerings, and that hasn’t stopped Cisco from investing resources into VCE, but this will?  Also, with VMware having a very small investment in VCE compared to Cisco and EMC, is the implication that Cisco will forsake a proven, growing sales channel for their fastest growing product over something like this?  How about we all use our brains for a second, OK?

“I think [the VMware-Nicera deal] is the last nail in VCE’s coffin,” said Vanessa Alvarez, president of Alvarez Consulting.

I’d like for Ms. Alvarez to a) contribute more than 10 words to the topic, b) please enumerate the previous nails that didn’t happen to be the last of them and c) get the author to spell Nicira correctly.  Again, it’s great copy and I’m sure will be great advertising for her personal consulting group, but it’s meaningless.  Not only does Ms. Alvarez not get the chance to explain any part of her comment, the rest of the article doesn’t expound on the existence of those nails either.  Maybe Ms. Alvarez had some great insight, but the author certainly didn’t bother sharing it.

Even VCE reseller partners once jazzed about selling Vblocks are dropping or reconsidering the product. Their question: Why devote time and resources to something that even the parent companies seem to be forsaking?

“I feel both Cisco and EMC are hedging their bets with their own individual solutions and they will continue to support VCE when the customer truly wants a converged infrastructure solution and they are competing against an IBM, HP or Oracle all in one solution,” said Jack Kaiser, SVP of sales for Focus Technology Solutions, a Seabrook, N.H. tech reseller. Companies like Focus that partner with all three VCE founders, have to figure out what’s best for the customer, he said.

This is one of my favorite sections.  Let’s count the issues:

  1. The opening paragraph references VCE reseller partners that are dropping or reconsidering selling Vblocks, and then a quote is given from a small VAR based in New England that employs less than 50 people and who has never been a VCE reseller (that I can find).
  2. The quote from the VAR shows an obvious lack of understanding for what a Vblock is and how customers are adopting it.  The goal isn’t to be able to compete with other converged solutions (a market VCE pioneered, by the way), it’s to show customers how a manufactured, converged infrastructure can make IT easier than putting the pieces together yourself and allow them to focus on the applications and the business.
  3. Yes, it’s true that companies like Focus have to figure out what’s best for their customers.  That is, in fact, their entire reason for existence.  I think Matt Davis summed it up best here.  Bleeding hearts, indeed.

Some see the appointment of EMC’s Pat Gelsinger as CEO of VMware as a move by EMC CEO Joe Tucci to tighten his hold on VMware. (EMC owns an 80 percent stake in the virtualization leader.) It’s clear to them that EMC with VMware and now Nicira is on a collision course with Cisco that leaves VCE and Vblock vulnerable.

Some?  Who is some?  Are we still talking about VARs?  Are we talking about analysts?  Once more, something is pointed out as “clear” with absolutely no supporting content.  Please, if it’s so clear, include some of the rationale in your article! I would LOVE for someone, Ms. Darrow included, to explain to me how VMware is on a collision course with Cisco in any way that implicates the movement of significant revenue.  It’s been days now, and the most anyone can do is “could, would, should, might, maybe” but no one has actually drawn a straight line from the Nicira acquisition to this “collision” being discussed.

Whether that growth proves meets expectations and is enough to sustain multiple converged infrastructure bundles when more work is flowing to webscale data centers running commodity hardware, remains a very big question.

Finally, we have the denouement.  After two quick paragraphs that include quotes from VCE that were given to other people for other purposes, she throws in two final jabs: is the demonstrated growth enough to meet expectations (whose expectations we are left to figure out on our own) and can it survive with workloads moving to “webscale” data centers running commodity hardware.  There was obviously no research into the service providers who have chosen Vblock (or any other reference architecture, for that matter) over commodity hardware.  There was certainly no data or references to back up the assertions.  It was just thrown out there like the smelly, rotting fish it was, to dry in the sun.

facepalmThis isn’t news.  This is hardly commentary.  There was no effort, it appears, to reach out to VCE for comment or information.  With the lone exception that a VCE employee took the picture used in the story (which he had to point out in the comments), there doesn’t appear to have been any input provided by anyone associated with the company either directly or from the partner ranks.  As a summation of the general reaction to the article from inside my team: we thought GigaOm was better than this.  Today’s story about the Oracle acquisition of Xsigo being about SDN and then Barb’s mea culpa after commenters set the record straight certainly hasn’t changed anyone’s minds…

John Furrier said it best when he commented on Twitter “as enterprise tech gets more consumer like it will put pressure on reporters to actually know their shit & still report.” I don’t want to pick just on Barb and GigaOm, because there are so many “news” outlets that are doing the same that it has to seem like the normal way of things.  As a consumer, I can tell you it’s not.  If you want to provide analysis, you need to have more than a press release and Google.  You need to know the players, you need to know the nuances and you need to understand the relationships (public and private) to the point where you can not only see the dominos, you can tell which way they are leaning.  Short of that, you are just making noise and hurting everyone in the process.