Hyperconverged infrastructure (HCI) is a hot topic in the data center world right now. Unless you’ve been operating your data center in a cave for the last three years, you’ve likely been deluged with chatter about this latest addition to IT infrastructure — and with good reason. HCI takes the promise of converged infrastructure and advances it to its logical next step. By combining compute, virtualization, connectivity and storage in easily deployable appliances, and using intelligent software to combine these into pools of resources, HCI has greatly simplified the design, deployment, support and maintenance of infrastructure stacks.
The key to HCI’s success is simplicity. It frees up IT personnel from the complexity of a traditional three-tiered infrastructure and allows them to take a business-centric approach towards technology. However, in all the excitement of deploying this new architecture and its promise of simplification, it is easy to overlook some key aspects of HCI deployment that must be kept in mind.
The Right Fit
While HCI is a great fit for most workloads, it is not the panacea that some manufacturers might claim. There are still some workloads that this architecture cannot accommodate. For instance, any time there is an extremely high performance application, care should be utilized to ensure that the HCI solution will meet those needs. There are also those applications (ones that will require physical rather than virtualized servers) that may not be a fit for HCI. However, as HCI offerings evolve, and all-flash nodes keep coming down in price, we see almost all workloads being a good fit.
The network design is a critical point to consider when building an HCI solution. The network demands of an HCI solution are unique. As an example, there is increased east-west traffic that must be accounted for due to the data protection techniques utilized by most HCI offerings. Also, there are specific network settings that an HCI vendor may require that will need to be considered. In a lot of cases, it makes sense to isolate HCI infrastructure to their own top-of-rack (ToR) switches.
Existing infrastructure touchpoints should not be ignored when deploying HCI. Before deciding on a solution, vendors should ask themselves: How will this solution affect the current backup & recovery solution and replication to disaster recovery sites? Beyond the access layer, do the aggregation and core layers of our network need changes? If the replication standard for DR is changing because of HCI, how does that affect WAN traffic?
Keep It Simple
Simplification of the infrastructure with HCI will be useless unless the IT team processes, procedures and resources are optimized to take advantage of the simplification. Most of us have worked hard over the years to define organization structures, standards and Information Technology Infrastructure Library (ITIL) processes in our environments, and it’s easy to overlook redoing these when deploying HCI. It’s no good having the capability to deploy a VM in minutes if our change process does not evolve to benefit from it.
By no means is the above a comprehensive list of aspects to consider when deploying HCI. There are many organization-specific points to consider as well. Hyperconverged holds great promise to revolutionize IT. If designed and implemented properly, it can set up an IT organization to be effective partners with business rather than a hindrance to overcome.
For more information about hyperconverged infrastructure watch this video or set up a meeting with your Data Center Solutions Architect.
This blog post brought to you by: