In a modern marketplace it’s important that technology is agile, flexible and scalable. In order to respond to these demands, enterprises need to consider hyperconverged infrastructure.
For years, the client/server/network architecture served organisations well – it was simple. As the number of applications increased, however, server utilisation became an issue. Provisioning could also take a long time. In some cases, a programmer might have to wait up to six weeks for a server to be provisioned with the right OS, applications and tools so they could do two or three days’ development work.
In today’s business environment where innovation, agility and speed to market are critical, such lead times are unacceptable.
Server virtualisation the first step
This eventually led to the introduction of server virtualisation, running multiple applications and workloads on the one machine. With the exception of some legacy systems, almost every workload can be virtualised and virtualised servers with a single pool of storage became the standard model of data centre.
Deploying this sort of infrastructure is quite complex and vendors such as IBM, NetApp, EMC, Cisco and Lenovo began doing it for their customers by selling converged systems. Basically virtualised servers, storage and networking in a 42U rack, these systems greatly reduced the complexity of deploying virtualised systems.
The concept has been so successful that traditional infrastructure sales are flat while the converged systems market was worth US$6 billion in 2014 with a 33 percent compound annual growth rate (CAGR).
Beyond convergence, hyperconvergence
The IT industry doesn’t stay still for long and two years down the track the hyperconverged system is the ‘new thing’. While converged systems included everything in a 42U rack, the hyperconverged system can include virtualised storage, servers and networking in a single appliance in a 2U box. Scale is achieved by adding another appliance, virtualising the storage and adding it to the network.
By combining the hyperconverged data centre with software-defined networking (SDN), organisations are getting ever closer to the holy grail of the software-defined data centre.
Reducing the burden of maintaining the IT environment
Another way of looking at hyperconverged infrastructure is to think of it as a private cloud data centre. Organisations deploying such infrastructure will become more agile, flexible and responsive with superior risk management and more predictable running costs. Because everything is virtualised, IT infrastructure management is greatly simplified and IT personnel can do it when they need to rather than when it will cause minimum disruption to the business. It also makes it easier to use outsourced infrastructure providers.
The potential of the hyperconverged datacentre is enormous. But unlike some previousdisruptive technologies that required a complete ‘rip-and-replace’, organisations can run new hyperconverged infrastructure in tandem with old infrastructure, which is just retired as it reaches the end of its life.