What do virtualization executives think about 2009? A VMBlog.com Series Exclusive.
Contributed by Andrew Barnes, Senior Vice President of Corporate Development for Neverfail
Virtualization. It’s a word that is unavoidable and seems to encompass everything from solving the world’s global warming issues to enabling all high availability issues and perhaps even being a cure for the common cold! But will the world be a truly virtual place in the future?
If you listen to the hype you may think so, but cast your mind back to the dotcom era for some perspective. Back then, there was to be only one type of business model: e-business. But has this come to fruition some 8 years on? Partially. Today, we have a business model that embraces physical and e-business platforms. Instead of e-business for e-business’ sake, it’s about the best technology to drive business success.
We’re facing a similar situation for virtualization, as I was reminded at Gartner’s Data Center Conference in early December. Virtualization has been around for decades on mainframes, yet the discussion there has not been about virtualization because it is the topic of the day. Instead, it’s about how to deliver business value. The time has come to stop seeing virtualization as a panacea and recognize it for what it is – an exciting tool that will contribute to the delivery of business-focused IT services.
A seemingly major benefit to virtualization is the ability to deliver high availability (HA). But let’s think about that for a while. HA to major virtualization vendors is all about rebooting the physical hosts and restarting virtual machines (VMs) in the event of a physical hardware issue. This is a necessity if you are putting all of your eggs in the proverbial (virtualization) basket. But in reality, it’s much more likely that a business service will be interrupted because an application in a VM has an issue. Perhaps Windows services have failed or a complete application failed after a new patch installation. Physical server HA driven by virtualization won’t help, so if you promised this, then beware. The investment in new hardware, shared storage and disruptive migrations may mean that HA is still a long way off. And if the business requirement was HA, where does that leave you?
A common misconception is that the various motion/migration tools available on virtual platforms are the solution to application availability. While that may be mostly true for planned maintenance, it’s not true for dealing with the unexpected. After all, if an application is “sick” in one virtual machine and it is migrated, it will still be sick, and business will still stop.
At that same Gartner conference there were two further, rather pertinent virtualization discussions --resource conflict and testing. In today’s highly unpredictable economic climate one thing is very predictable: competitive pressure will place more demand on IT systems and staff, and any disruption to IT systems – at any time of the day – will be unacceptable. With global customers and a mobile workforce, 24x7 availability will become the norm, at least for critical collaboration and information systems such as e-mail and databases. Even today, finding time for backups on dedicated servers can be a challenge while finding backup windows for multiple virtualized applications on the same physical server can be impossible due to resource conflicts. Virtualizing applications at the wrong time can lead to less availability.
So what can be done to take advantage of virtualization while delivering true application availability? It’s always best to first look at the business and then technology needs. When looking at availability, it’s important to have an independent, up-to-date, continuously replicated copy of the data supporting the applications. This means redundancy of data, not a single copy on shared storage. Using virtualization combined with host-based replication to provide that redundancy makes a lot of sense. Having a data availability hub means that multiple applications can be protected with multiple VMs on one physical server, without the risk and disruption of physical migration of the primary applications. But to actually eliminate downtime, or at least get it to less than a minute, data availability alone isn’t sufficient. Using the availability hub to monitor the health and status of applications means that if issues occur the applications can be failed over – immediately. And if the right architecture is in place, with the right tools, this can mean that users are not disrupted at all, including physical-to-virtual migration and application failures.
As with “e-business,” the best high availability solution is a blend of physical and virtual platforms. It leverages virtual machines together as an availability hub and leaves the physical machine untouched. This is a better approach to application-level HA (and even disaster recovery for that matter). It also can pave the way for ultimately virtualizing those critical applications, after testing, and testing again, in a non-disruptive way. But that’s a topic for next time…
About Andrew Barnes
Andrew Barnes is Senior Vice President of Corporate Development for Neverfail.
He joined Neverfail in March 2007 bringing extensive experience in marketing, product management, and pre-sales through his 25 years in the software industry. In his new role, he will be responsible for Neverfail’s branding, marketing, product management and web presence. Barnes most recently served as VP of Marketing for a European-based software company and has held a variety of senior positions with companies such as KVS, Sun and Platinum Technology. Previously, Barnes worked for three years as Global Director of Marketing for KVS Inc. where he was responsible for all aspects of marketing and grew the customer base tenfold until the company’s acquisition by Veritas (now Symantec).
Prior to KVS, Barnes was Northern Europe Marketing Manager for iPlanet and Product Marketing Manager for Forte Software (acquired by Sun Microsystems). In addition, Barnes served as the European Product Manager for Platinum Technology where he led a multinational team responsible for the launch and sales enablement of Platinum Infrastructure management products across Europe.
For more information, visit http://www.neverfailgroup.com/ .