Legacy infrastructure, that is, the traditional method of building a data center around “point” products (servers, storage for capacity, storage for speed, backup appliances, recovery and orchestration software, etc)hasn’t changed much in the past twenty years.
However, the problems associated with managing these types of environments as new and faster products are added, coupled with the ever-increasing amount of data being retained for analytics & security, is now reaching the breaking point for most IT managers, who can no longer can they staff and maintain this amount of independent parts. This is just one of the many forces behind HCI’s rapid adoption and the creation of both Converged and HyperConverged Systems in the marketplace. Here’s a brief summary of both.
There are similarities and differences between these infrastructures(legacy and HCI) that you should be aware of. Converged infrastructure relies on appliances that arrive pre-integrated and pre-defined for specific purposes and ready to deploy. It’s also a unified solution that simplifies the data center by offering compute, storage, networking and server virtualization resources in a single platform, which guarantees the components will work in unison. In addition to arriving preconfigured, the hardware is capable of scaling to power larger (or smaller) workloads without sacrificing reliability.
Hyperconverged infrastructure is comprised of modular solutions enabled by software-defined components. Unlike converged infrastructures where the compute, storage, and network capabilities are housed in one appliance, hyperconverged adds a layer of software that provides centralized automation, management, and control. This solution removes some of the barriers inherent in converged infrastructure by allowing IT organizations to rapidly configure and deploy resources to any workload that calls for them. The result is saving your IT staff’s time and adding much needed scalability.
While HyperConverged is on the fast track, Gartner predicts hyperconverged will have amazing growth at 79% within five years, but why haven’t more businesses considered moving to this new type of infrastructure?
For one, there are issues around scale that the software and network gurus simply haven’t mastered. With one of the many selling points of HCI being redundancy and almost instentanious data replication between nodes for data protection comes the overhead of having to pay the penalty of latency and demands on the tools used (memory, cpu, ssd’s & disks) to ingest the data. I predict that within 12 months, this issue is a non-factor, and Gartner’s estimates will seem cautious at best.
For now, most large scale applications will continue to run on all-flash arrays, leveraging large farms of blades for massive compute resources, but in the absence of these monolithic farms, HCI is simply the way of the future, and should start proving to the traditionalists that it, too, can play in the Enterprise landscape. For more information, read our blog on HCI, or leave a comment below in the feedback forum.