This is the first in a series of articles that explore the impact of public cloud service outages, such as the recent AWS failure. In this series of articles, Leidos UK/Europe Chief Cloud Architect Lee Benning explains how you can work with clients to reduce the risk of using public cloud services for their business critical applications and data.
Over the years, I’ve worked with many large corporations in various industry sectors. Differences aside, there was one common trend with all these companies: the way they adopted new technology at an uncomfortable pace, often driven by the demands of a strategy or need to cut costs. The first signs of this were in the adoption of virtualization and private clouds.
I was often engaged by desperate CxOs and IT directors or managers facing situations where technology had spun out of control and taken on a life of its own. The ‘promised land’ that would deliver rapid deployments of infrastructure, crystal-clear costs and tame responsive control, had instead become a series of plagues of VM sprawl and black holes of virtualization consuming vast computing and storage resources. All while administrative and support staff shrugged their shoulders and stared on in bewilderment.
Dollars, CPU cycles and storage terabytes evaporated in ever-increasing totals and complaints piled up about poor responsiveness and delays to key initiatives. Virtualization, automation and self-service were supposed to have the opposite effect. The board demanded answers. Why? What was going wrong? How could it be fixed?
I see similar trends in the adoption of public cloud services. As economics drive clients to seek savings and bring agility to corporate IT, it dictates a path ever more to the promises made by cloud. Whilst some have learned from their past mistakes, many customers still adopt the cloud thinking that, by default, it will save money and speed things up. They expect cost of ownership graphs to arc down, indicating the anticipated savings, but instead they see the cost curve climb up the dollar axis.
Cloud hype is driven by promises of lower costs of ownership, flexibility and almost infinite elasticity, which are some of the key attractions touted by advertising. These are, for the most part, true claims, if and only if you understand the cloud and how it functions. Public cloud is fundamentally different to how you run your corporate-owned service. Three common mistakes that occur are:
- Failing to adapt applications and services to the cloud. Many adopt a migration approach that relocates VMs from the internal infrastructure to the cloud ‘as is.’
- A failure to appreciate the requirements of your own service back to your users and on-site environment.
- Thinking that once it’s in the cloud, it will always be available because clouds never fail.
The first two points are related in that they can be overcome as cloud adoption issues by fully understanding your application architecture. Points two and three are related in that they underline the importance of understanding the impact of outages and poor performance on your clients’ users and, potentially, on their business. Both of these issues can be rectified by ensuring that your clients have a clear vision of what the cloud should bring to their operation. Like any other project, there will be a set of objectives which should be measurable and followed up on after the project completes to evaluate whether the change was justified and delivered on its goals.
Adopting the cloud then, is not as simple as it seems. It’s a complex process and requires a large degree of discovery, analysis and planning. Once your clients have embarked on the journey, it’s important to work with them to make adoption as painless as possible. The other half of this is to ensure that the way the service gets deployed does not bring about unplanned outages through poorly designed cloud services. This will immediately undermine the value of its adoption. Our focus here then is the second part of the coin, which is to ensure that disruptions to business are avoided through proper planning and design.
Read my next post in the series to learn what we advise and how your clients can work with us to avoid being part of the headlines next time a major service goes down.
Lee is the Chief Cloud Architect for Leidos UK/Europe.More Content by Lee Benning