Today’s IT provisioning needs to happen fast — not the slow, methodical weeks or months of classic back-room systems, or even fast-tracked hours-to-days, but often within minutes. Basically faster than a system admin can type a few characters or make a few mouse clicks, particularly if this is only one of the many tasks they have.
Developers and testers need faster provisioning to support today’s compressed idea-to-development-to-deployment cycle, and the tighter-focused, quick-iterating Agile methods. Where applications used to take months, even years to develop and deploy, and non-security updates happened only monthly, quarterly, or yearly, many companies with “user-facing” apps may want to create and push out changes several times a day, even hourly, on an ongoing basis.
Virtualization lets system instances be provisioned fast enough to accommodate these compressed cycles. The limiting factor is the availability of appropriate management tools.
The same changes in speed and frequency also apply to de-provisioning — “spinning down” software instances. Today, IT needs to turn things off just as quickly and easily as turning them on — to replace an instance with a new version, to free up resources occupied by instances created to perform a specific, finished task, or to support a now-waning demand surge.
This need for rapid spin-up/spin-down doesn’t just apply to operating systems (and, by extension, to virtual machines and containers). It also applies to other components of the software stack — including database management systems (DBMS).
Supporting Database-as-a-Service for a cloud computing environment like OpenStack increasingly calls for faster, flexible lifecycle management, for several reasons:
- Developers — and therefore, also testers and operators — are increasingly likely to use multiple, different DBMS’s, selected for each task’s requirements. (See Why Are There So Many Database Management Systems?) This increases admins’ needs for more capable tools – to create multiple configuration templates, to click-and-select provisioning choices.
- The demand for DBMS provisioning of instance spin-up and spin-down will either require growing amounts of system administration, taking admin time away from other tasks, or need response times faster than human reflexes can achieve. A DBaaS that includes automation can address both these challenges.
- DBMS instances need more than just database spin-up and spin-down. They need performance monitoring. They may need backups. They may need careful configuration to work with each other on a larger, distributed task. A full-featured DBaaS will let admins (or end users) do these procedures more quickly and easily.
- DBMS’s need to respond quickly to handle failover. Slow database spin-up isn’t acceptable here, of course. DBaaS tools can expedite and even automate the spin-up for faster service restores.
Rapid growth also demands efficient spin-up. A start-up that begins with only a dozen or servers’ worth of instances can easily ramp to a thousand — or ten thousand — within a year. But that only works if the software stack is as “elastic” — able to grow that quickly and easily. That includes whatever DBMS’s that are part of your company’s cloud stack.
This is where a DBaaS and cloud management tools come into play – to help your developers, operators, testers and system administrators create and configure, deploy, monitor, manage, and de-provision DBMS instances in your cloud. With the right tools, your IT people can wrangle DBMS instance starts and stops quickly and efficiently to meet your organization’s needs.
Photo Credit: “Spinning Windmill” by Nick Page via Flickr