VMBlog: 2017 Prediction Series
Enterprise transformation will continue apace in 2017, and for good reasons. Businesses need to update existing systems to take advantage of the agility that modern architectures enable – not for the sake of technology itself, but to better align people, process, and technology with business strategy. Agile methodologies allow this transformation to occur much faster than traditional software development and architecture lifecycles, as they enable rapid releases, learning, and iteration. Adoption of public or private cloud technology can greatly accelerate transformation efforts, as it allows on-demand scaling, programmatic control and deployment of infrastructure, automation of deployments using blue/green paradigms, and rapid provisioning of QA and Dev environments when needed.
So why is cloud adoption, while robust, still in pilot or trial stages for so many enterprises?
While there’s no shortage of cloud resources available (AWS, Softlayer, Azure, etc.), there is a lack of staff experienced with cloud technology. The set of Developers and Operations people that are experienced both with cloud services and with building/running critical applications is fairly small. Existing applications that assume high levels of reliability and consistent performance often need significant architectural changes to run in a cloud, which is done by separating processes into scalable units, connected by queuing systems. But failure to use experienced architects can lead to troublesome failure modes if, for example, latency to the queuing system increases and all threads retrieving queue items are stuck waiting.
Another key reason for hesitancy about headfirst cloud adoption is cost.
Yes, running in a public cloud provides all sorts of agility advantages, and cost savings for short lived or dynamic workloads. Other cost savings like not having to deal with ISP contracts, power provisioning, and network equipment service contracts, add up as well. But there are going to be a lot of workloads that are stable and ongoing (or increasing) workloads. Those workloads – especially if they require a lot of memory or performant disks – will be much cheaper to run in an enterprise’s existing datacenter. And no CIO wants to get his budget caught in cloud jail.
So what to do?
Enterprises need to invest in their staff so they gain expertise in cloud deployments. This will enable them to benefit from the agility of the cloud, and give them expertise in assessing when applications, that may have been through a few iterations in the cloud, should be brought back to the datacenter.
One way to free up staff so that they can gain these experiences is to focus on automating processes as much as possible. Push adoption of orchestration tools; replace checklist processes with build automation; embrace SaaS tools that free up datacenter resources (and more importantly eliminate server and application maintenance and backup); adopt an automated IT monitoring tool that provides insight into both cloud and on-premises deployments. The effort will help in increasing IT agility directly and the increased efficiency will allow more learning of cloud technologies.
Check out the full article on VMblog here.