As an industry it’s been quite a journey when it comes to cloud computing. What started as a safe environment for testing and development quickly morphed into infrastructure as a service. Economic drivers then saw everything shift to the cloud, but many companies and research organisations discovered not every application is well suited to that environment.
Most of the companies I work with agree the cloud is fit for purpose for the vast majority of workloads and business applications currently in use. What I am hearing from these companies are the questions they are asking as they move towards a hybrid strategy. What makes the most economical sense to keep in the cloud? Where can I get a competitive advantage by keeping certain workloads in the cloud and what should I do myself?
In many organisations critical assets need to be close to key resources in order to produce the data and answers that enable them to maximise their business outputs. These are big questions with no single set of answers that work for everyone. Each company has to make these determinations for themselves.
Organisations are turning to high performance computing (HPC), artificial intelligence, machine learning, and more to move their business forward. These workloads require large amounts of data to be relevant. They also require a greater level of control within the environment - of the data, operating system, hardware, connectivity, etc - to ensure the workloads are as fine tuned as possible. It's a bit like buying a brand new Range Rover and discovering you can’t put it into all-wheel drive. A high powered machine throttled by functionality.
Bare metal enables the level of control needed for these environments. High performance benefits from the HPC cluster in one location, as opposed to using virtualised servers in a cloud environment. Genuine parallel processing can take place within a very managed environment. Access and connectivity to the public cloud then allows data to be dispersed within wider database networks. Finding the right colocation partner who can enable you to create this environment is key.
The first time I visited Kao Data’s campus in Harlow, I was impressed with what I saw. Kao Data has made it its mission to develop and operate high performance data centres for advanced computing inspired by hyperscale and that is exactly what I witnessed. It is on par, if not a bit better, than the data centres I worked with during my time at Microsoft.
The expectation with a hyperscale cloud is that you will be getting the latest technology with green credentials, it will be highly efficient and massively secure. With Kao Data, you have an equivalent experience, plus the benefit of controlling your own bare metal. You know exactly where it is and what it’s doing. It’s processing your HPC workloads 24/7 and transacting with data that is a millimetre away. It also provides excellent scalability and flexibility for future growth. You may not know what you need tomorrow or next year, but with bare metal you can change it as needed. It’s the best of both worlds.
The challenge most organisations face is that they are only as good as their worst process, and unfortunately, IT can be a bottleneck. One person’s view of the cloud is likely to be completely different to another. The CSO, CIO and DevOps teams will all have different pain thresholds and business objectives. The need to control your own destiny will lend itself to a hybrid environment for the next decade. Make sure you have the right partners in place to help you do so.
Mark Bedford
Mark is a Non-Executive Director at Kao Data, Board Advisor and Consultant, with over 35 years of experience within IT including Microsoft (GM Microsoft Solutions), VmWare and SAP.