There are a lot of good products out there, to be sure. And data growth shows little sign of slowing. But companies have ratcheted back their spending on Big Iron arrays, driving a lot of profit out of the projections of storage platform marketers.
We saw it coming a mile back. We forecasted the "perfect storm" -- a commingling of a slowing economy and the resulting "do more with less" business mantra, mixed with the commoditization of hardware -- in this column a couple of years ago. Comments by former CEOs from leading hardware vendors, including EMC Corp.'s Mike Ruettgers, that software was becoming the only differentiator between arrays that were essentially just boxes of Seagate hard disks, only reinforced our view. And in the case of IBM, EMC and Hewlett-Packard Co., the redefinition of themselves as software and services companies sealed the deal.
Now, the array vendors are finding that the only segment of the market showing real growth is the small to medium-sized business sector.
Even the SAN vendors are clawing desperately to maintain some market by dumbing down adapters and reducing prices on their gear overall in a desperate effort to sell their overpriced wares to SMB consumers who have no need for them. It is little surprise to us that so many former Fibre Channel bigots have now embraced iSCSI in a big way. With their lower price point and capability to leverage the cost-efficiencies of Ethernet and TCP/IP switching, iSCSI is poised for big growth in littler companies.
What is bad news for vendors is potentially good news for consumers in the short term. However, longer term, we are all going to suffer if a few steps aren't taken today. Here is my recipe for putting the blush on the storage rose.
1. Deconstruct the array. Get rid of all the overblown controllers and expensive proprietary software and externalize these functions to run on less expensive devices outside the box. Offload things like point-in-time mirror splitting from your most expensive disk to ancillary SATA platforms using third party software, such as Revivio's Time Addressable Storage. Replace expensive data replication software that forces you to use the same vendor's product at both ends of the network with vendor-neutral software like SOFTEK's Replicator (TDMF for open systems). Instead of buying proprietary and controller-focused data tagging mechanisms that commit you to one vendor's platform forever, use platform agnostic software from companies like Avamar Technologies to handle reference data migration and tracking. The list could go on and on, but the point remains: Deconstructing Big Iron will drive cost out of arrays and decimate vendor lock-in strategies.
2. Buy a vendor-neutral management package first, then only buy storage products that it supports. Turn the tables on the Big Iron guys who "gift" their management software to you in the hopes that you will stick with their products solely. Think outside the box, literally. A less product-centric management toolset (think Computer Associates, Tivoli, Softek, Tek-Tools, etc.) will enable you to capitalize on a broader range of technology products and price points while at the same time ensuring the manageability of your infrastructure going forward, which is key to cost containment.
3. Purpose-build your storage. Think about your applications and what they require in terms of the disposition, handling and accessibility of their data. Forget the one-size-fits-all marketecture promulgated by SAN vendors over the past few years. SAN is just an interconnect. The important thing is selecting the right disk, the right interface, and the right interconnect to support the application. If you need scalability at the lowest possible cost, check out hybrids like Spectra Logic's RXT or Adaptec/SNAP's latest platforms. For small to medium-sized shops, look at combo disk/tape units like Breece Hill's iStora 4000, which my test lab guys are hammering on today and discovering performance characteristics that are making them into believers.
4. Think disaster recovery. Build it in to your product choices. Redundant power supplies are a must, but also look for external connections that will facilitate data replication to anybody's gear. Develop a high availability/recovery strategy, then make vendors explain how their products will work with it. Make disaster recovery a vetting criterion in product selection.
With the current state of storage, consumers are in the driver's seat. We can either take control or we can continue to do what we have been programmed to do by lots of vendor marketing dollars: Go for the proprietary solution because it is the safe and easy play.
Many companies are doing the latter. Brand name gear can be gotten at rock bottom prices these days, and that helps to sweeten the decision. What is missed by most of the folks who opt for this path, however, is that the advantages of a homogeneous solution today may translate to significant cost, not to mention missed opportunity, tomorrow. I have been doing a bit of traveling lately with Randy Chalfant, a technology architect for StorageTek and one of the smartest guys in the business. Randy has some impressive financial projections pertaining to the longer term costs of going with the brand name, monolithic, Big Iron vendors that make it clear that this is a financially questionable decision.
The alternative, that of purpose-building storage from the ground up, seems to add complexity at first glance. And nobody wants complex these days. However, with management framework conformity as a selection criterion, the vicissitudes of complex, multi-vendor solutions may well be offset by the lower cost of the solution both at the time of acquisition and over its productive life.
Something to think about.
About the author: Jon William Toigo has authored hundreds of articles on storage and technology along with his monthly SearchStorage.com "Toigo's Take on Storage" expert column and backup/recovery feature. He is also a frequent site contributor on the subjects of storage management, disaster recovery and enterprise storage. Toigo has authored a number of storage books, including Disaster recovery planning: Preparing for the unthinkable, 3/e. For detailed information on the nine parts of a full-fledged DR plan, see Jon's Web site.
This was first published in August 2004