Blog by Daniel Lamorena: Linux and High Availability - the best of both worlds?
by Daniel Lamorena - Director Storage and Availability Marketing
Use of Linux in the data centre is continuing to grow. Looking at the year-on-year analyst figures from the likes of IDC, Linux server usage is increasing a few percentage points a year – this is largely down to Linux taking a bite out of legacy UNIX installations (Windows is holding steady). Organisations are not migrating away form proprietary platforms for altruistic reasons – making a shift can result in quite considerable savings in terms of both hardware and software.
This doesn’t tell the whole picture, however. While Linux is pretty solid at a kernel level, it doesn’t offer as comprehensive facilities as commercial operating systems when it comes to more advanced features such as High Availability. This is less about functionality, and more to do with ease of configuration – customers tell us that the features available ‘out of the box’ require additional effort so they deliver what is required, for example in terms of writing scripts and tweaking capabilities to fit with specific requirements and work with specific storage platforms and applications.
Given that we’re talking about High Availability, these are elements that organisations want to get absolutely right. As one of our customers, HD Supply told us, “When internal network problems occurred, a multi-node Linux cluster for a mission-critical Oracle database didn’t fail over as expected.” As well as adding a configuration overhead, then, the margin for error is still too great in case of failure.
Unsurprisingly perhaps, we’re seeing growth in demand for our Storage Foundation products for Linux that far outstrips that of the operating system as a whole. This proven technology works across all major software and hardware, ensuring cross-platform support and ensuring a level of familiarity – the latter should not be under-estimated, given the need to minimise the potential for configuration errors.
In response to the demand we’re continuing to extend the capabilities of Storage Foundation High Availability, for example with Virtual Business Services in the 6.0 release, which enables complex application dependencies to be defined and built into restart behaviour in case of major failure. We also continue to work with Linux vendors to deliver the most effective solutions (for example, with Red Hat, around disaster recovery). Put simply, failure is not an option for our customers, so nor is it for us.
In enterprise IT it is important to see the full picture, whatever the platform. While it is never a good thing to be complacent, we're comfortable with the fact that our facilities can reduce costs when things go well, at the same time reducing risks in case of disaster. If you'd like any further information, please visit here.