One of the best parts of my job is hearing how the IT leaders in our customer base are succeeding in delivering non-stop operations for their companies’ most vital apps. They serve many different industries and sit in companies with a few hundred to 10s of 1000s of employees, but despite those differences, they all have one problem – there’s simply no time for downtime.
Across these many conversations with these innovative CIOs and architects, along with findings from Gartner’s recent data center conference, a picture of the evolving data center is crystalizing. To deliver non-stop operations, they’ve adopted new architectures and new processes – they’re building data centers at webscale. Let’s look at this change in more detail.
The changing definition of a data center.
What is a data center? For many organizations, it’s no longer a physical building. Instead, it’s about delivering compute resources. For some of our customers, regulatory restrictions mean the old-fashioned approach of company-owned facilities and devices must continue. But for other businesses, public cloud and hybrid have taken over – IT must simply deliver compute capacity on demand.
Why webscale is crucial.
The operational efficiency of applying webscale technologies and methodologies has proven to be needed at all but the smallest scale operations. IT organizations of any reasonable size are struggling with the volume problem – keeping up with increased service demands just doesn’t scale if you’re relying on old models for building out software and running systems. Modular architectures, reusable code, and automation are all crucial for keeping pace. According to Gartner research, IT leaders cite Webscale benefits as faster time to market, higher quality systems, lower costs, more agility, and the ability to stimulate change in the organization.[1]
The fundamental principle of webscale – resiliency vs. recovery.
One of the hardest parts of running webscale data centers is shifting the mindset from recovery to resiliency. As systems get bigger, errors and failures become more common – the key is for element to fail but not take down the overall system. Our customer Mouser Electronics has designed for just such an expectation – architecting the data tier for resiliency saved the company’s online transactions during a recent complete hardware failure of the primary database.
How to build webscale.
The webscale principle of resiliency must also apply to your people. The hardest shift is developing a culture that expects failure and solves it quickly vs. moves slowly and avoids any mistakes. Your systems will also need to change. Design your apps to run in multiple data centers. That architecture can present a challenge at the data tier – ScaleArc has solved that challenge for many enterprises, including Microsoft IT. Standardize and automate as many processes as possible – you can no longer afford to have “guild-level artisans” crafting custom systems, chides Gartner analysts. Simplify your hardware infrastructure as well – standardize on fewer choices and deliver more consistency. It makes procuring, sparing, maintaining, and swapping gear much easier.
Webscale at the data tier.
I recently discussed the challenges of building webscale at the data tier. Applying an abstraction layer like ScaleArc to the data tier delivers these webscale benefits to one of the hardest parts of your infrastructure – the database. As you’re architecting your systems and your staff to run at webscale, know that at least one part of the challenge has been made easier, with ScaleArc software.
[1] Gartner research note: Principles and Practices of DevOps?