Cisco's future in storage - Alternative and contrarian options
In RE: to the first three posts….. I believe the author was either being facetious when referring to Cisco as the "Storage King” OR it’s a really cryptic way of saying Cisco MUST acquire EMC which would literally make them the storage king (more on that below).
Here's my somewhat contrarian thoughts on the topic. Cisco has to go cheap or differentiate themselves with whatever they decide to do (or NOT to do) with storage. And deciding NOT to do do anything (with respect to acquiring a storage company) now may be a wise alternative, assuming they instead invest and position themselves for what ‘storage’ will be 18 - 36 months from now. There's no sense in Cisco going after any of the AFA outfits that are left, since none of them are either A) differentiated or B) have the architectural foundation to become differentiated without 5 years of R&D. All of the pure storage plays have either gone public or in the process of being acquired (Solid Fire + NetApp). I’d have to imagine that Cisco’s board of directors can’t be too enthusiastic having their new CEO go after a storage company given their $400M+ acquisition of Whiptail and subsequent write off only ~24 months later.
BUT, why not just punt on doing any external storage acquisitions? At least for now anyhow. Cisco not acquiring a traditional or AFA storage company is only a risk to growing future earnings. Traditional persistence that’s been a function of traditional Tier 1, external storage arrays for the past coupe of decades,,,,, well that isn't exactly a high growth market. Future architectures are moving toward design centers that put the data right next to the CPU (or the CPU next to the data).
So if the 18 - 24 month “industry analysts” crystal ball predict customers moving toward “flash and trash” (as the article refers to it, I prefer the term “Hot Edge / Cold Edge” myself), these types of architectures and deployment patterns require a HIGH SPEED LOW LATENCY PCIe GEN 3 TYPE of FABRIC to connect a boatload of cores to a bunch of shared NVRAM or a dense rack scale type of NAND (I’m not talking about AFA type of NAND)……. I’m thinking Cisco may have a few engineers that could incorporate something like that into the UCS ecosystem. And for the cold edge? Who cares it’s a commodity, though Cisco could easily solve that with a cloud gateway appliance to S3 or whoever. It’s not like gizmo’s that connect an on-prem data center to a public cloud provider are in short supply. Anyone remember Cisco’s WAAS and WAFS gizmo’s from a decade ago?
Speaking of Cisco gizmo’s past with respect to storage acquisition alternatives, couldn't Cisco leverage the engineer and code that already exists in their Storage Service Module (SSM) to make a controller based virtual storage type of solution? I’m thinking along the lines of something that resembles IBM’s SVC in architecture whereby the storage controller is divorced from the backend storage arrays, allowing them to capitalize on the part of storage that’s valuable to customers and have better margins than storage capacity. Think about data services like local or remote replication. Or the things like storage abstraction, the pooling of disparate arrays to automate and orchestrate storage operations and service levels via auto-tiering type of functionality. This route would give customers flexibility and choice of OSM, JBOD, AFA, etc and prevent Cisco from having to get into the traditional storage array market where commoditization is underway.
But as for where storage architectures are headed, maybe Cisco should consider picking up someone or developing something like PernixData or Infino, assuming it can be designed to run on platforms beyond VMware, (bare metal Linux OS’, containers, Hyper-V, OpenStack, etc).
As for storage solutions sold to customers for their VMware environments, I think it’s an opportune time for software based storage solutions that are decomposed & distributed across compute (the same compute customers have to buy anyhow) that leverage EFD in the host. Why now? The maturity, growing acceptance of VSAN coupled with it’s beefier maximum limitations should help it gain traction in the larger enterprises. And then there’s VVOLs right around the phased GA release corner (no really). Point being, we’re seeing VMware and the greater VMware ecosystem offer functionally that are alternatives to and was only provided in the million dollar monolithic array (at scale). Granted these solutions still don’t scale in comparison to a Tier 1 array but EFD’s are getting bigger and I think next year we’ll see 3D TLC offer much larger densities that we’ve seen from prior EFD’s.
Any of these options would give Cisco their own assets which are not tied to their soon to be divorced polygamy J.V. ex-realative(s) within VCE. They can't afford to take their eye off of the ACI on ASIC vs NSX on commodity war. If Cisco doesn't pick up a reputable storage solution, they’ll live, albeit with growth having to come from somewhere else (which how much of it comes from storage now?). If however, they loose the network virtualization war, well that may start a long death spiral as customers move toward commodity network hardware.
Which leads to “Dropping the following bomb”:
If Cisco feels it needs to acquire it’s way into the storage market to complete their stack and grow earnings into the future, they should go after the right storage company. Not another little man-n-egger with a few hundred people and some IP that’s a fork of ZFS. Wait for it, wait for it, here it comes…… Cisco MAY still be interested in acquiring EMC. By no means is the Dell offer a done deal. Yes the mandatory ‘go shop’ period is over, BUT EMC board of directors still has to consider any offers that may be better for the share holders, who still get to vote on it. Plus the penalty for breaking off the Dell deal is only $2.5B. Considering they’d be offering something north of $70B, $2.5B is nothing. Nor is it uncommon. As a matter of fact EMC paid NetApp a breakup fee of $57M, when they outbid NetApp to acquire DataDomain for $2.1B. Which on a percentage basis is right around the same amount relative to the deal price if the plug is pulled on the Dell / EMC deal.
If Cisco acquired EMC it’d be a two birds - one stone type of deal. First, from a strategic perspective, they get a storage outfit, engineers, IP and a war chest of storage solutions that are #1 in every market they play in. So they’d get street cred across all segments of primary (AFA & hybrid) and secondary storage and emerging storage (SDS, Scale out NAS, VTL, SRM (the other SRM), etc, etc, etc. Secondly, it’s a wise tactical move and would secure their ASIC committed future with ACI by “combining” ACI and NSX, thus protecting their core business for another half decade at least.
Cisco’s has over $60B in cash, cash equivalents and short term investments. Their total assets are north of $110B. So they have the money to do a deal without funding it via a tracker stock, that no shareholder would prefer to all cash or real Cisco stock, as tracker stocks typically trade at a 10% discount to the real stock do to not having any voting rights or other shareholder privileges that come from owning the actual A-Shares. Tracker stocks are very really used now a days. So if I’m on the Cisco board or shareholder, I have to wonder what can you do with all that money that’s a better alternative to acquiring EMC? Cisco’s market cap is ~$140B, so they could do a tax friendly all stock deal or some stock + some cash to sweeten the deal for shareholders while minimizing the overall tax burden.
Beyond solving ACI vs NSX, think about how Cisco + EMC would solve the VCE JV for Cisco and give them a platform / direction for their hyper-converged future (which is missing now). I mean what can possibly happen with all of UCS sales that are sold as a vBlock from VCE, if the Dell / EMC thing goes through? Maybe not right out of the gate but shortly there after?
Anyhow, there’s my contrarian views of Cisco’s future in storage. BTW…. I have no first hand knowledge of anything that isn’t public and this is just one simple storage guy’s opinion.
Biting the hand that feeds IT
