Building Storage, Part III
Acquisitions, new releases, value-added hype --- what can we believe and who is offering something new?
Continuing to identify the decision points inherent in building storage infrastructure, we take a closer look at the software elements of storage infrastructure.
There has recently been a good deal of activity in the storage software space—last month’s acquisition by EMC of RSA Security, HP’s acquisition of Mercury Interactive, and CA’s buyout of XOsoft, an infrastructure high-availability and continuous-data-protection software company. Add to this announcements by Pillar Data Systems that they were retooling their marketing focus on their software components and a similar announcement by LeftHand Networks, and you can see the increased importance of software in the stories of storage vendors.
EMC has been talking the talk of a “kinder, gentler” software player for over two years, though the mantra is punctuated by the occasional release of yet another proprietary and monolithic hardware platform. The announcement that it was acquiring security veteran RSA Security raised few eyebrows, since it seemed to be a natural fit with their effort to be regarded as the premiere provider of regulatory-compliant information repositories, if not another alternative in the operating system stack wars currently dominated by Microsoft, Oracle, and various peddlers of open source OS (such as Sun with Solaris). Reading between the lines (and sometimes reading their words as expressed by CEO Tucci and others), one can see EMC endeavoring to become a credible rival to soup-to-nuts vendors such as IBM, rather than remaining simply a storage player.
VMware is the jewel in the EMC software crown. Regardless of what you might think of its maker, just about everyone is wild about VMware, which delivers an excellent capability to virtualize and consolidate servers. Other components in the growing software stack, such as Documentum, elicit a less generally positive response from IT managers—especially those for whom the idea of a workflow-based content management system holds little appeal.
RSA is a stalwart in the security business, which is both a plus and a minus from the standpoint of selling into EMC’s traditionally storage-focused clientele. On the plus side, the RSA acquisition suggests that EMC is serious about becoming a full service company, including solutions for critical storage requirements in its offerings. On the negative side, RSA had recently begun garnering some criticism from folks who have soured on Public Key Cryptography (PKC), a technology for identity protection that relied on public and private tokens or keys to authenticate and secure messages. RSA owned the patent until 2000, and built its entire company around it. However, managing a proliferating number of keys has become the bugaboo of PKC technology in general, and in recent years has cast a shadow over any enthusiasm about the method.
Still, EMC will now have bragging rights that it has a synergistic security play to go along with its storage story and to compliment its server virtualization and electronic content management plays. It will be interesting to see whether it resonates with consumers, some of whom are complaining of the lock-ins and high prices already charged by the company with its more traditional storage wares.
EMC is the archetype of the company that has been promising to replace a hardware-only focus with a software focus. Last week, I could have sworn I was hearing the same message from Pillar Data Systems, which markets to EMC’s clientele: larger medium-sized and enterprise customers. Their big news was their intention to re-tune their marketing messages to favor their software elements, which they call AxiomONE.
Speaking with Russ Kennedy, Senior Director of Marketing, he noted that the company had been shipping product (the Axiom 500, a mid-range SATA array with a clustering architecture and chameleon-like support for both NAS and FC fabric connection) for about a year. Kennedy said that the terminology of the company’s campaigns had focused on hardware rather than on software, and that this was now about to change.
“Hardware is [a] commodity. We are transitioning to software—switching gears to emphasize the software components that make our product unique.” Kennedy told me.
AxiomONE is a collection of software tools that come with the Axiom product. “Not a software stack” per se, to quote Kennedy, AxiomONE delivers the ability to “tier storage based on performance levels” and quality-of-service guarantees required by the data itself. Other tools include managed replication, backup, and something unique called SecureWorm FS—a file-system-level, write-once-read-many (WORM) technology that is interesting to customers concerned with data nonrepudiability.
The Pillar story sounds a lot like EMC because it only works on their platform. Says Kennedy, “We deliver solid performance. A good story. But I’m not saying that we will eventually open the hardware platform.” This statement, at least, merits kudos for honesty. Contrast this with EMC, which has promised software-only versions of such products as its Centera content-addressable storage system since it first started selling the platform, but has never delivered.
One problem with proprietary software, however, is the exposure it creates for customers to the disappearance of the vendor from the market. Pillar has about 100 paying customers, according to Kennedy. Not bad for a two-year-old company that has actually shipped product for little over a year.
However, the company owes its store currently to Oracle boss Larry Ellison, who has proven a bit mercurial in prior hardware investments such as Network Computer. Kennedy insists that there is no reason to be concerned. Ellison made an impromptu visit to the shop when they were celebrating their 100th customer and said some agreeable things about the importance of storage and the firm belief he held about the need for an alternative to the big-iron players who currently dominate the enterprise market.
More credible, perhaps, is the decision of LeftHand Networks, a provider of iSCSI solutions, to open its software to support third-party gear. The company’s secret sauce SAN/iQ software for nodal clustering, virtualization, network RAID, replication, snapshotting, and thin provisioning is much beloved by the company’s current customer base. We have certainly never heard a discouraging word from those who have deployed LeftHand wares, and CEO Bill Chambers and CTO John Spiers are two of the most credible and genuine folks in this industry.
Bottom line: if they say that they are going to open things up to other SATA array vendors, we believe them. Furthermore, we look forward to seeing how they do in their new path to market.
Computer Associates and the OpenView group at HP have always been about hardware agnostic management. Both have made important acquisitions of late that compliment their portfolios. HP’s acquisition of Mercury Interactive helps the company substantiate its claims to be moving further up the stack to application performance-based management. However, OpenView is still less an integrated management solution than a common launch pad for numerous, minimally integrated, software tools.
We contrast them with CA and the difference is night and day. Perhaps the sole benefit of being under the SEC’s microscope for a couple of years, CA turned inward in 2003 and did some back-end code-level integration of many of its storage management wares, and it shows. Of late, they have been on a buying jag that has included an e-mail archive solution, a records-management solution, and now, with the acquisition of XOsoft, a recovery-management solution.
In XOsoft, CA adds a complementary application and server failover capability along with a fairly robust continuous-data-protection story to its data protection and high availability story, which heretofore centered on ARCserve, its backup product line. Bob Davis, senior vice president and general manager for CA’s BrightStor product family, chatted with me recently to put the deal in context.
Interestingly, he emphasized XOsoft’s high-availability server/storage failover component, WANSyncHA, to a much greater extent in the conversation than he did the Enterprise Rewinder CDP component—despite the storage-centric nature of the latter. Without reading anything into this, it seemed clear from Davis’ remarks that the real value add that he saw in the XOsoft purchase was the flexible, user-definable, application-failover, strategy-building tools that distinguish the product. This is where integration with CA’s other wares must happen to enable greater automation of both data management and data protection. Davis clearly gets this point.
All in all, it has been a busy summer in the storage software space. It will be interesting to see if all the work produces results. Traditionally, consumers have failed to prioritize management as a component of their storage hardware acquisitions. These vendors, to a one, are banking on the notion that current concerns with disaster recovery and regulatory compliance will help promote their wares.
Your comments are welcomed at firstname.lastname@example.org.