An Alternative Approach to HCI

Hyperconverged Infrastructures claim to be a simpler alternative to a traditional architecture with a dedicated compute and a dedicated storage tier. The primary source of the simplification is the elimination of a dedicated storage tier. Instead, HCI integrates storage into the architecture by aggregated storage inside the servers that make up the compute tier. As discussed in our previous blog, as HCI scales, balancing storage resources and guaranteeing performance becomes more complex. Instead of throwing out all of the shared storage lessons learned over the past two decades, what if the storage system was improved and optimized for working in a virtualized environment?

The Advantages of Dedicated Shared Storage

When compared to HCI a dedicated share storage system certainly has some advantages. First, the network and the storage systems compute are dedicated to the task of storage IO. It does not have to compete for attention with other higher level processes running in the environment. That means an application that spikes in compute need will not impact the storage system that supports it. Even if the spike drives increased demand for storage IO the dedicated storage system has processing power waiting for that task.

The Disadvantages of Dedicated Shared Storage

The problem with legacy dedicated shared storage is that the design of most came in an era where the standard application was a physical database dealing with structured data. That database ran on dedicated server hardware with dedicated network connections. There was also no real need to have multiple servers accessing that data, in fact, IT wanted quite the opposite; guarantees that no other application could access it.

Virtualization Further Challenged Dedicated Shared Storage

Virtualized environments are clustered environments by design. They want every server in the cluster to be able to access the other server’s data. In the virtualization use case, this cross server visibility is needed to enable virtual machine mobility.

Another challenge is, although the shared storage system had its own dedicated processing, making sure that certain applications were assured certain levels of performance was not a common capability. The lack of performance guarantees was made worse in a virtual sense since it was not the server that needed the guarantee it was the virtual machine within the server that did. However, as we’ve discussed, HCI does not provide good visibility nor guarantee performance.

Modernizing Shared Storage

Instead of throwing out the concept of shared storage it may make more sense to modernize shared storage for the virtualized use case. The first step is modernizing access; a file system semantic makes more sense compared to block storage for virtualized environments since virtual machines are essentially file images of servers. From a management perspective, file systems or NAS are considered by many to be simpler to manage than block-based systems.

A file system approach also provides the foundation for the storage system to gain deeper insight into the needs of each virtual machine, which includes VM level quality of services so specific VMs can be assured of the performance they require and other VM can be limited so they do not consume more resources than they should.

A key concern of HCI is resource efficiency as it scales. These systems can’t scale up to the maximum capabilities of the node before adding another. Most can’t add nodes, within the same cluster, of different types. A shared storage system with VM level visibility can create a loosely coupled cluster of a storage system where each member of the cluster can have different capabilities, and VMs can move automatically based on analytics to the most appropriate node based on IO and capacity requirements.

StorageSwiss Take

HCI solves some legitimate problems and simplifies some virtualized environments. But as HCI scales, storage management and network design become increasingly critical. Assuring consistent and specific VM performance becomes almost impossible. Instead of starting all over with HCI, it may make more sense for the organization to look a modernized shared storage designed for the virtualized environment. A file system based approach that leverage VM level insights to assure required application performance outcomes may be the better alternative.

Sponsored by Tintri

George Crump is the Chief Product Strategist at StorONE. Prior to StorONE, George spent almost 14 years as the founder and lead analyst at Storage Switzerland, which StorONE acquired in March 2020. In his spare time, he continues to write blogs on Storage Switzerland to educate IT professionals on all aspects of data center storage. He is the primary contributor to Storage Switzerland and is a heavily sought-after public speaker. With over 30 years of experience designing storage solutions for data centers across the US, he has seen the birth of such technologies as RAID, NAS, SAN, Virtualization, Cloud, and Enterprise Flash. Prior to founding Storage Switzerland, he was CTO at one of the nation's largest storage integrators where he was in charge of technology testing, integration, and product selection.

Tagged with: , , , , , , , , , ,
Posted in Blog

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 21,785 other followers
Blog Stats
%d bloggers like this: