For CIOs with businesses running on SQL Server 2005, time is running out–they will no longer be supported by Microsoft as of April 12, 2016. Yet, just a few months ago, IDC estimated there were 800,000 servers running SQL Server 2005. It’s no surprise, because upgrading to SQL Server 2012 or SQL Server 2014 can cost a large business millions of dollars, clearly requiring careful planning and consideration. Many businesses delay upgrades even in the face of material long-term benefits because of the inevitable short-term disruption. "Don’t fix what’s not broken" is the mantra.
The fact is that with SQL Server 2005’s EOL, sticking with an unsupported database for business critical applications is not an option. For savvy CIOs, this is an opportunity to push through typical organizational gridlock and modernize their business critical infrastructure stack. They can leverage this compelling event and short-term disruption to not only refresh their database deployment, but also refresh the infrastructure stack to reap long-term benefits for the business. So, what should CIOs look to do? CIOs should look to pursue strategic initiatives in three key areas to deliver long-term value to their business–virtualization, performance, and analytics.
Virtualize the Last Mile
The foundational initiative CIOs should pursue is the virtualization of their SQL server deployments. As virtualization has matured over the years, it’s now possible to virtualize some of the most demanding databases in terms of size and performance. Virtualizing SQL server pays huge dividends in the long-term. Higher consolidation through virtualization means cost savings through higher hardware utilization, and lower software licensing costs since SQL server is licensed on a per-socket basis. Virtualizing SQL server deployments also enable a much more agile datacenter, leading to lower operational expenses. Businesses can stay up and running even in the event of server failure through failover protection at the hypervisor level, and also leverage replication and orchestration for easy disaster recovery during a site-wide outage. When it comes to the eventual hardware refresh down the road, the live migrations enabled through virtualization mean no disruption to the business. CIOs also need to prepare for a long-term hybrid or public cloud strategy, and virtualization is the on-ramp.
Pursue a Solid-state Performance Strategy
A faster business is a more competitive one, and high performance is a must-have for mission critical databases that support revenue generating or productivity-essential applications. CIOs should look to a strategy that can both deliver the performance they need today as well as easily scale as needs change in the future. Flash is the biggest trend in storage today, which generally delivers great performance for databases. Deploying flash in a storage array helps boost performance, but for the most demanding databases, CIOs should also look to leverage flash and RAM in the server where it is closest to the processor for minimal latency. Upgrading to the latest and greatest SQL Server 2014 provides in-memory capabilities that can dramatically speed up database performance, though additional planning is required, as the database has to be re-architected to take advantage of this capability. Another option is to use technologies that can pool flash and RAM at an infrastructure level in the server, which can essentially give in-memory computing capabilities to any SQL Server database (such as SQL Server 2012), while eliminating the need to re-design the database. The best long-term strategy for performance is usually to deploy flash both in the storage array, as well as in the server to serve as an elastic low-cost insurance policy that delivers maximum performance as needs change.
Leverage Infrastructure Analytics
Businesses today are data driven, and IT needs to keep up. CIOs should put the right 'big data' analytics in place so that they have the insights they need to make the right decisions. Without an infrastructure analytics solution in a dynamic datacenter, it becomes very difficult to answer questions such as "How do I fix a performance problem for a critical application?" The problem could be at the SQL database layer, at the virtualization layer, due to the network, or due to some idiosyncrasy with the storage media (flash isn’t great for large block writes for example). With analytics, the answer can be understood at a glance, along with how to fix the issue. No more guesswork and war-room scenarios when the inevitable critical issue pops up.
SQL Server databases are often the workhorse that a business depends on, and for that reason, they are often untouched unless there is a compelling reason to do so. With the SQL 2005 end-of-life, CIOs have the reason they need to take advantage of short-term disruption to finally implement strategic initiatives that can contribute both to the top and bottom line of the business in the long-term.