* More from Karp's crystal fishbowl
Some people use a crystal ball, but if your name is Karp, you of course use a crystal fishbowl to look into the future. Last time in this newsletter, I pointed to two major trends, vendor consolidation and the return of the service providers, which are likely to take hold in 2007. Peering past the angelfish, loaches and miniature piranha, here are some of the other things I see for the New Year.
Trend #3: Management software gets closer to managing systems rather than objects.
Historically, IT has managed hardware objects – in storage, this has meant host bus adapters (HBA), arrays, NAS boxes and so forth. Much of this has been because, until the last few years, we had a situation where hardware assets were managed, for the most part, by software provided by the hardware vendors themselves. At large sites this resulted in a series of ad hoc management tools, one for the HP devices, one for the EMC boxes, one for the IBM stuff, and so forth – hardly an economical approach to management as it resulted in redundant skill sets and HP specialists, IBM specialists, etc. This of course mirrors the server world, where we are used to seeing HP-UX specialists, Linux specialists, ands so forth. Lately, with standards-based SRM, some management capabilities have been extended to include storage devices from multiple vendors, helping us move from managing objects to (almost) managing domains.
What’s the next step? 2007 will see the first really credible efforts to manage storage within the context of the entire IT system. This means looking at, understanding and analyzing the interplay between storage, systems, networks and security, a topic I’ve been carping about for years now. Happily, some of the players are finally beginning to get it right, offering us cross-domain management capabilities (not all the domains, but correlating data from any number greater than one is a good start) that move us a bit further in the direction – not of managed objects – but of a single managed system. An increasing number of vendors, including the traditional “framework” vendors but also including many much smaller companies – are coming up with techniques to do this.
Trend #4: Fibre Channel and SCSI disks will be replaced by SATA on the network.
I actually saw this happening last year, but networked storage – both SAN and NAS – will see a big up tick in the use of these cheaper drives this year. And why not? These drives will certainly fail at some point – as will all the others – but my friends in the data recovery business tell me that their business (which relies on disk failures) indicates the failure rate for SATA seems to be tracking its growth in the marketplace. Depending on the storage environment, cheaper may indeed prove to be more than just “good enough,” it may turn out to be better. Thus, SATA may even start to appear on tier one devices.
This will be very interesting to watch. Expect the most severe implications to be in the Fibre Channel segment; SCSI will maintain its strength in direct-attach situations, but the vast majority of Fibre Channel devices are networked. Expect the mix of networked Fibre Channel and Fibre Channel-connected ATA to shift significantly in favor of the much cheaper ATA.
Trend #5: The Red Sox will win the pennant.
They will trounce the Yankees; after that, who cares about the World Series? More on that this autumn.
