Interesting times! Cisco and Whiptail. One opinion..

Disclaimer – I’ve said it before, will say it again, this blog is my blog, not an EMC blog.   It’s inevitable that since I am a leader at EMC that I will have my bias, the flip side (hopefully for the readers) is that the perspective comes with a unique seat at the intersection of a lot of customers, a lot of engineering, and a lot of what happens in senior levels in a lot of places.

Today, Cisco announced their intent to acquire Whiptail.   Cisco’s PR is here.  Their first blog post is here.

Flash is changing the landscape.

For the last 6 years, EMC has investing in every way we can (on the server software stack, on the all-flash array market, and in the hybrid array market) – it’s a case of “self-disrupt, now!”   

We’ve called out the fact that Flash is one of the mega-disruptors from a pure technology standpoint - the others being: 1) low cost, high performance compute/memory/networking; 2) virtualization/abstraction in general; 3) cloud automation and consumption models; 4) convergence blending of compute/network/storage.  

Like has happened before, these are all “fundamental tech disruptions” that underpin the mega-trends that people see of Mobile, Social, Cloud, Big Data.

As I said last week at VMworld:

“Flash is changing everything about storage [editors note – what I mean here wasn’t just “arrays”, but any form of persisting data].   Sounds like hyperbole?  It's not.  Sometimes people don't "get it".  Flash is like someone coming along and saying "I have a new car on the market.  It's pretty awesome.   It fits in your pocket, but can still get you around.  It's gets 1000 mpg.   Oh, and it can go 10,000 mph."   That would be a pretty radical inflection point, no?

Again, my analogy may sound hyperbolic, but it's no exaggeration.   That's flash relative to magnetic media.   Yes, the $/GB is still higher, but getting closer and closer every day.   For some workloads, $/GB pales compared to $/IOps or the relative value of a few hundred microseconds per IO.  Sometimes density and power dominate.”

Anytime that a new fundamental “building block” technology arrives on the market (flash was invented in 1980 by Toshiba), it causes waves of startups, waves of innovation, and waves of industry disruption.  That’s what’s so awesome about high-tech IMO!

So – in that context, what do I think about this news today?   If you’re interested in one perspective, read on!

I’ve been at a startup or two.   in fact 3… 2 that blew up (part of the fun!), and the one that EMC acquired – was a app-focused storage stack (think like Tintri) that was a software-only storage scale out stack (think like Lefthand/ScaleIO/VSAN).

Startups are interesting in a different way than a large portfolio company.   Every startup has at it’s core a tech idea and a business/market idea.   Flash storage startups start with the “flash gives an opportunity to disrupt the $60B storage market, and I’ll do it by leveraging flash THIS way…”.

Like stars they burn bright, and then burn out, fade slowly or go super-nova.   Often it doesn’t happen in one cycle, but within a single startup, they repurpose what their IP in different ways as their business ebbs and flows – continuing the star analogy, they “go super-nova and then the shockwave triggers new star formation”.

In the flash startup ecosystem – it’s brutal.

Each startup is like a “star” - all in various parts of their lifecycles. 

• Some are gas clouds condensing to hit that gravity/fusion point where they quickly flash into light (many I follow that are in stealth mode).   
• Some of the early “red giants” are burning out (look closely at Violin). 
• Some are in that “supernova/shockwave” and are working re-invent themselves (look closely at FusionIO).  
• Some of the darlings whose star shone brilliantly from just last year are competing to get attention over the new cool kids on the block whose star is starting to burn as they compete for almost the same use case (look at Nutanix and Simplivity).   

Whiptail was one of the early players in the market with some very interesting IP, but never got to “escape velocity”. This may sound like I’m being pejorative – I’m not (and hey, they just got bought for $415M, so that highlights what some value them for).

This is the crucible of startup innovation – and it burns HOT.   It’s this heat that generates so much goodness for customers and the market as a whole.

Also, none of the startups live in a vacuum only looking at customers – ignoring everything around them.   As you would fully expect, they are constantly compared to their peers, they constantly circulate in the VC community, they are constantly under evaluation by the large players.  

That ecosystem also includes factoring in the “ripples” of other things constantly going on in the marketplace (so in this case, you need to think of think of the Pure Storage funding round 2 weeks ago, Virident acquisition by Western Digital last week as factors, certainly in terms of valuation).

If you think through that lens:

• Every server vendor is looking at “how can I apply flash” in what I do?
• How fast will there be (if indeed there will be) commoditization of the hardware (think intel on motherboard)?  Should I instead try to innovate at the controller – can I do that?
• What is the software stack to win?   Is there IP I can buy or build to get there?
• How does this apply in the different workloads (think in-memory databases vs. VSAN and storing virtual machines) and markets (think high frequency trading in the financial vertical vs. the SME/SMB) I play in?   How will those change over time?
• Every networking vendor is looking at “how can I apply flash” in what I do?
• Is there an intersection of flash on a crazy-low-latency network/top-of-rack?  (think EMC Project Thunder from last year, that now lives on in application in various ways in future EMC things)
• Is there an intersection of SDN and SDS upon us I can leverage?   (think of this thing that I showed here)
• Every storage vendor is looking at “how can I apply flash” in what I do?
• How do I build next-generation hybrids – as these continue to form the broadest market?  (think of last week’s Speed2Lead launch of the next-gen VNX)
• How do I build next-generation all-flash arrays (build, buy – think XtremIO, etc) – as these will inevitably disrupt at least some of the current marketplace.
• How do I look at the server market which is disrupting the “shared storage array” market as a whole (think ScaleIO, VSAN and more).
• Lastly, everyone is looking at the fact that the lines are getting very blurry in the world of convergence – and many of the players play in one, two, or many of these markets.

I’m very fortunate to be at a smart company, filled with very smart people – and have demonstrated a proven ability to self-disrupt through different several vehicles.

Cisco is a very smart company, filled with smart people – and has demonstrated that they can use various vehicles to self-disrupt (most notably the team behind Andiamo, Nuova Systems, and now Insieme) .  

Look at today’s news through that lens, and ask yourself these two questions:

Q: "If you were Cisco, and you were wanting to throw down the gloves, and play in the existing array and current broader storage market, and fight furiously with your partners – who would you buy?”

A: What is your answer?  Mine is “I can think of a few, but they aren’t Whiptail”.

Q: “Conversely, if you wanted to apply the disruptive technology that is flash into your existing strong market of servers and network, and work to expand a new category of persistence – who would you buy… mostly for their IP?"

A: What is your answer?  Mine is “I can think of a few, one of them would be Whiptail”. 

I fully expect some analysts to jump to the conclusion that Cisco is trying to answer the first question I asked above.  

I get that.  Everyone likes to be polemic, and stir up a fight – after all, it gets more eyeballs :-)  

I think that wiser heads that reflect for some time tend to be more correct.  

I would expect those people to echo the analysis we saw after VMworld on the VMware/Cisco SDN topic – and their question (and answer) will be more like my 2nd question/answer.   If you think about the SDN discussion post VMworld, the initial responses were aggressive, I would say even defensive.   Eventually, more calm rational voice came outacross the broader community (I liked this one, and this one).  I think those more calm rational voices are more analagous, and more instructive to today’s news.  

My personal opinion?  

I think Cisco wants (as they state in their release and the blog) to leverage the disruption of flash and some of the Whiptail IP to try to leapfrog their competitors in the server and networking space – and in doing so bringing a new type of persistence layer to the server and into the fabric.  

Perhaps over longer time periods, this is of course COULD be disruptive to the industry (unless others get there first or in different ways).

I would expect Cisco to be initially focusing the IP at the more of the very high end spectrum of IO (think of my high-frequency trading example) that is about tight server/network coupling with some persistence – rather than the broad marketplace as a whole.

Against that backdrop… Cisco and EMC are great partners.   This manifests itself thousands of times every day in:

1. Our joint customers (who are legion – thank you!)
2. Our joint partnership (VCE and Vblock – thank you customers!)
3. Our joint go-to-market with our distributors and channel partners (VSPEX – has accelerated to be the largest “reference architecture” solution in the marketplace after a late start – thank you partners!).  

Of course, I think it would be disingenuous to suggest that there is no interesting stuff in the speculation people will make around today’s news.

Conversely, I can think of a million ways we’ll be able to extend the successful elements of our partnership, just in the same way that VMware and Cisco can and will partner around SDN.   It’s what customers want. 

I can imagine (and in fact am working on) some really cool ScaleIO + Cisco opportunities in VSPEX, and I know the VCE crew is working on how to build extreme forms of Vblocks using the latest + upcoming technologies from VMware, EMC and Cisco to continue to deliver the most compelling converged platform on the market.   I think that innovation in all forms helps customers.  

“May we all live in interesting times” – has been true, and continues to be true!

Comments

[Keith Norbie]

Well said Chad. The interesting winner here may become the flash software players like Pernix.

[B]

How does this effect the next gen VBlock release and the relationship EMC and VMware have with Cisco? Doesn't sound good.

[Marcelo Fandino]

Agree with you that we are living in interesting times. Not only EMC will do smart moves. We need to focus on strengthening our partnership with Cisco adding value to the ecosystem. While we would be able to do this, it will work.

[Glenn Grabowski]

Hey Chad. I'm starting to use the "building block" analogy, too, as I discuss how to build a SDDC. Great minds think alike. haha.
Lines are getting blurred though as EMC tries to push it's server-side flash technology while Cisco acquires IP to possibly do the same(assuming they aren't looking at WhipTail to be an "array"). How will VCE choose which technology is used in a vBlock? Time will tell but it isn't exactly a positive event.

[JP Morgenthal]

Chad,

It seems as though your take overlooks the architectural implications of Whiptail in the UCS architecture. If I understand correctly, Whiptail is a memory architecture. When linked into UCS I would expect it to provide addressable RAM for applications to run in. ExtremeIO is a persistence layer derived from Flash. Together I would expect unbelievable performance for moving data from at-rest to in-memory, but yet not overlap in capability. Someday, Cisco could determine to do a single level store where addressable RAM an persistent storage are one in the same, but for now it seems a pure Flash Vblock would require both Whiptail and ScaleIO.

What's your thoughts?

[Mjbrender]

Disclaimer - EMC'er here. Still, I was as curious as everyone else on this one.

I really appreciate the clarity you bring to this conversation Chad. The combination of startup experience and logic makes everyone more informed by the end of the post.