A great deal has already been written and said about the announcement of the new EMC VPLEX earlier this week. But why anyone considers the VPLEX itself so revolutionary is a bit perplexing to me. Companies like Datacore, FalconStor, HDS, IBM, NetApp and even EMC (think Invista) have enabled this architecture for at least a decade.
The real news this past week out of EMC World is not that EMC has decoupled its VMAX or Symmetrix controller heads from its back end disk drives, added some bells and whistles to it and called it “VPLEX“. The big news in my mind is that this decoupling puts the storage industry on notice that EMC has officially begun its transformation from a disk vendor into a provider of storage intelligence.
VPLEX itself was announced during the keynote presentation of EMC Information Infrastructure Products President and Chief Operating Officer Pat Gelsinger amid thunderous music and appearing onstage out of a cloud of smoke. Then as Gelsinger went into what new possibilities the VPLEX enabled and an EMC engineer came onstage to demonstrate a live data center fail over from one site to another, those who were relatively new to the storage space (less than five years exposure), thought these new capabilities were the best thing since sliced bread.
Having been around the storage block for a few years, I found myself somewhat unimpressed with the announcement as a whole since I have been evangelizing about this architecture for nearly a decade myself. But as I listened to Gelsinger and watched the VPLEX demos, I did find myself becoming excited about what this means for the storage industry as a whole.
Say what you will about EMC, it is the big dog in the data storage industry and now that it is throwing its weight behind storage virtualization (or “The Private Cloud” as it now prefers to call it) in general and VPLEX specifically, this has some positive ramifications.
First, it ratifies the concept of network-based storage virtualization. While EMC may argue until it is blue in the face that the VPLEX is controller-based storage virtualization, when EMC CEO Joe Tucci gets on stage and says that the VPLEX will be given the ability to virtualize other vendor’s storage systems, it is with all apologies to Tucci a network-based storage virtualization architecture – or at least a variation thereof. If anything, the storage provider that should feel most justified by the VPLEX announcement is HDS as it appears EMC was in part looking over its shoulder when it designed the VPLEX.
Second, existing EMC VMAX and Symmetrix users can bring forward their existing scripts and interoperability matrices. One of the largest obstacles to storage virtualization in existing enterprise EMC shops was the inability to bring their existing scripts and interoperability matrices forward into a virtualized storage environment. They had spent much of the last 20 years developing interfaces, methodologies and scripts to work with existing EMC Symmetrix products and all of that work and best practices are not easily or lightly abandoned.
Now that VPLEX is available, this is no longer an obstacle as the VPLEX will work with their existing scripts and interoperability matrices and they can bring those forward into a VPLEX environment. In this respect, VPLEX is to the storage virtualization world what VMware is to the server virtualization world. It creates a standard platform on which to deliver storage services.
It is important to note that only some of this functionality will be available immediately. For instance, out of the gate, VPLEX will only virtualize other EMCs products (CLARiiON, Symmetrix and VMAX).
But what is equally important to note is that the storage intelligence that end-users have built into their existing EMC storage environments (user logins, LUN mappings, LUN groups, SRDF scripts, etc.) can now be brought forward into EMC’s version of the private storage cloud. Granted, this will result in some form of storage vendor lock-in but most users already operate under that premise so at least it will be easier to manage.
Third, it will accelerate the adoption of the private storage cloud and storage sales for both EMC and its competitors. As much as all of EMC’s competitors like to rail on EMC, the VPLEX announcement is EMC’s ratification of the viability storage virtualization and the private storage cloud. This is good for everyone – EMC and its competitors. So while the term “storage virtualization” is now politically incorrect and the term “private cloud” is in vogue, EMC has just made it OK for everyone to talk about the private storage cloud and promote it because EMC has publicly and officially endorsed it.
Fourth, users will pay more and be locked in but they will likely be happier with the end result. In a breakfast meeting that I had with Peter Mehta, the CEO of SANpulse Technologies during EMC World, he said that EMC Services derives 40% of its revenue from data migrations. In my humble opinion, that is insane. Storage networks were supposed to create shared storage pools and make it easier to share used and unused storage capacity between applications, not more difficult. Yet without the right form of storage virtualization, externally attached storage becomes incredibly difficult to manage.
The VPLEX will not eliminate the need for data migrations but it should eventually make them easier to execute. While users will likely need to be prepared for one final painful migration into the VPLEX cloud, once that is done, data migrations should be much easier to accomplish going forward. Much of the complexity associated with data migrations will be minimized with more control afforded to end users to execute on them.
Fifth, the VPLEX architecture looks to lend itself to organizations centralizing the management of data across their entire enterprise. The VPLEX is clearly intended for enterprise shops with large data centers but it is available in two node configurations. So in just looking at the VPLEX from a bird’s eye view and the configuration options that are theoretically possible, there appears to be nothing to prevent organizations from using the VPLEX to eventually bring all data throughout the organization under its management. In so doing, they can create a centralized and consolidated approach to manage their data and their storage.
It is this last point that may be the biggest news coming out of EMC World. It is not that EMC announced the VPLEX or that it has officially endorsed the private cloud. Rather it is that EMC has finally demonstrated that it is serious about becoming a storage intelligence company. VPLEX demonstrates that EMC now understands that the value in round, brown spinning disk is no longer there. Rather EMC’s new value add is to bring all storage capacity under the management of the VPLEX regardless of the storage provider is where the real long term play for it is.
Now that EMC is actively promoting that message, it is not only good news and bodes well for EMC, it is good news and bodes well for the entire storage industry.