Of course it does – just not for everything. A growing number of enterprises, particularly in the financial services sector, are realising that applying 99.99999% SLAs to all workloads isn’t necessarily the most cost-effective way to run IT infrastructure. As developers become more agile, businesses should consider how they can manage SLAs in a way that’s as flexible as their cloud infrastructure: not just seeking to lower costs, but to be able to “scale” availability up and down as their needs shift and evolve.
Let’s say you’re running a low-level workload to test a new service – part of your new DevOps strategy, perhaps. The service doesn’t need extremely high levels of uptime because you’re still in the testing and iteration phase: in fact, your developers want to see when the service fails to identify potential areas for improvement. But let’s say the service experiences far larger demand than expected during public beta, so much so that it starts to hit performance issues or outages. Your uptime requirements have just changed and you need a way to quickly “scale” your SLAs to keep up.
At the same time, there will always be mission-critical services which require the highest uptime possible. You recognise these need best-of-breed infrastructure, without compromise: a VCE Vblock, for example, rather than an off-the-shelf DIY configuration. Yet these systems need to integrate seamlessly with your lower-availability workloads – ideally running on the same platform. How do you cover both ends of the spectrum without breaking IT operations into siloes?
The solution lies in a managed-service approach to converged infrastructure. Our customers increasingly tell us they want full control over how they consume IT resources, particularly when it comes to costs and responsibility for day-to-day operations. At VCE, we both provide best-of-breed hardware (like Vblocks and VxRacks) and look after it as a fully managed offering, giving our customers the security of SLAs while allowing them to vary the levels of these agreements in an agile way.
This is very different to your traditional model of IT outsourcing, where any change to your SLAs or other aspects of your configuration typically incurs an (let’s face it) unreasonably high cost. Those outsourcing models are in real danger of going extinct as more and more businesses adopt “utility” models for IT infrastructure. Although they may (but not always) cost more, the flexibility and control of spend compared to CapEx investments is increasingly worth it.
Is your business taking a multi-level approach to SLAs and uptime?