The future of storage
- 12 January, 2006 11:05
Predicting the future is perilous business indeed. However, as an industry analyst one of the sad facts of life that I have had to face is the reality that it is awfully hard to find clients willing to pay me to predict the past. This is an unfortunate situation, but it is, alas, the way of things.
So, keeping in mind the 59th Rule of Acquisition ("Free advice is seldom cheap"), here is some commentary on what I expect to see happening in the storage business during the next four quarters:
The marketing sizzle around continuous data protection (CDP) will continue to grow, and so will the steak. An increasing number of companies will add event-driven backup and archiving as a supplement to their time-based offerings. This means that users will be able to recover data with greatly improved granularity than has previously been possible. The first appearances of CDP have already occurred, primarily in support of Exchange. Look for the technology to be extended to Oracle and other databases in 2006.
Large and medium sites looking to maximize the value they get from their storage assets will continue to move tiered storage out onto the IT room floor, deploying SATA-based devices where SCSI and Fibre Channel were previously the only choices. Larger sites will begin serious consideration of data grids within their own firewalls.
It is worthy of note that a side effect of the interest in grids is likely to be the re-emergence of storage service providers (SSP), who will find the on-demand asset allocation (and pay-as-you-go billing approach) of grids to be right in line with the way their customers will want to use their services. Of course, IT shops will have to be willing to let someone else handle a significant part of their corporate data, but the economics of this model are likely to be compelling and that will help get many IT managers past their initial reluctance to let data go outside their own firewall. Grids are coming folks. And SSPs are re-emerging.
As storage environments grow in size and complexity, expect to see even greater evidence of what I have been talking about for a year now - the increased need for cross-domain manageability in complex environments. This demands much more than simply managing networks, storage, applications and so forth through a common console - it also means having software that can cross-correlate the events that happen on each of those subsets of the overall IT system and then understand the root causes of the events that occur. The need to understand the problems, and not just the symptoms, is becoming paramount.