As hyper-converged technologies continue to mature, organizations are increasingly considering how to use them...
in their environments. Hyper-converged architecture is probably best known for its role in virtual desktop infrastructure, but it is evolving to support a wider array of use cases.
Startup organizations that choose to run virtual machines (VMs) locally in their own data centers naturally gravitate toward hyper-converged infrastructure. However, this technology is beginning to gain traction in well-established data centers as well.
Some organizations are gradually replacing aging virtualization hosts with new hyper-converged infrastructure. Even the most cutting-edge hardware will eventually become obsolete, and many organizations have adjusted their hardware end-of-life plans to include the purchase of a hyper-converged system in lieu of a hardware upgrade.
Although not as common, some organizations are replacing current generation production hardware with hyper-converged infrastructure. This rip-and-replace approach to hardware upgrades is usually discouraged because it means replacing perfectly good hardware with something new. In spite of the cost, some organizations have found they can save on expenses by transitioning to a hyper-converged architecture even though their server hardware has not yet become outdated.
A move to hyper-converged technology might reduce costs due to a higher VM density, lower latency and support costs, and greatly reduced storage costs.
Pay-as-you-grow cost model appeals to branch offices
Regardless of whether an organization chooses to adopt a hyper-converged architecture within its data center, many enterprises are discovering that the technology is a good fit for branch offices. Branch offices have always presented a unique set of challenges because they are isolated, typically lack a dedicated IT staff, do not have the resources associated with a traditional data center and usually have minuscule IT budgets.
Hyper-converged infrastructure tends to be on the more affordable end of the spectrum when it comes to enterprise hardware, and it is designed to be scalable. An organization can create a small-scale deployment in a branch office and grow that deployment on an as-needed basis by adding nodes. This pay-as-you-grow cost model allows costs to be kept in check.
The compact physical design and turnkey simplicity of hyper-converged systems allows them to be deployed in remote environments with minimal effort. More importantly, many hyper-converged products support remote management, which means the entire system can be managed and maintained without IT presence in the branch office.
Perhaps the best reason for adopting a hyper-converged infrastructure in a branch office is that it provides an easy way to achieve high availability for virtual servers. Any data storage administrator who has ever had to support a branch office knows how disruptive it can be when a server fails. Providing a low-cost method of achieving high availability means things will run more smoothly and the IT staff will have to make fewer trips to the remote location.
Arun Taneja, founder and consulting analyst at Taneja Group, explains the distinction between hyper-converged and converged systems. He also discusses products for hyper-converged architecture and predicts the direction these products will take.
Hyper-convergence helps test and dev
Hyper-converged architecture is also being used in development and test environments. This may seem counterintuitive given the cost of the required hardware, but costs can be kept in check by using a minimal number of nodes. The deployment can always be scaled out later by adding nodes.
A hyper-converged system is a good choice for use in dev and test environments for two reasons:
- It is self-contained. As such, organizations can use hyper-converged system hardware to create development or test environments that are completely isolated from the production environment. Administrators can be assured that none of the work done in the dev and test environment will have the potential to adversely affect the production environment.
- It is specifically designed for use with virtualization. Development and test environments tend to rely heavily on VMs. Lab environments must typically limit the number of VMs that are powered on at any given time to prevent everything from slowing to a crawl. Because a hyper-converged architecture is based around true, enterprise-class hardware that is specifically designed for virtualization, it can handle dev and test workloads that more realistically reflect that of a production environment.
Backup and disaster recovery emerge as use cases
Some organizations use hyper-converged architecture for disaster recovery purposes. A big trend in IT over the last few years is the transition from traditional backups to virtualization-aware backup applications that feature instant recovery capabilities. These types of applications work by replicating VMs to a backup server. If a virtual machine were to fail, an administrator can initiate an instant recovery, which causes the VM to be mounted and run directly from the backup while the recovery operation takes place in the background.
Instant recovery capabilities have revolutionized the backup and recovery process. For these capabilities to be effective, the backup and recovery environment must have adequate hardware capabilities. An under-powered backup server will run VMs more slowly than those same virtual machines might run in a production environment. This performance hit -- which is commonly related to storage that cannot deliver a sufficient number of IOPS -- is made worse by ongoing backups and the background recovery process, both of which cause additional IOPS. It is also possible that a recovery operation may require multiple VMs to be run from the backup server, exponentially compounding the server's load.
The key to making instant recovery work is to use the proper hardware. A hyper-converged system is a good choice for backup and recovery environments because the storage has been specifically designed to support multiple VMs running at the same time.
Although hyper-converged architecture is most commonly used in virtual desktop infrastructure environments, it can be useful for other purposes. As it continues to mature, there is little doubt the technology will be used to handle a wide variety of mission-critical tasks throughout the data center.
Hyper-convergence for VDI is a good bet for scalability
There was a growth in hyper-converged technologies in 2015
Install hyper-converged architecture in tandem with existing infrastructure