Advanced Computing in the Age of AI | Tuesday, November 29, 2022

Software-Defined Storage and Hyperconverged Infrastructure: 3 Use Cases 

The data center has become associated with complexity as traditional storage is struggling to keep pace with workload demands. To address these challenges, industry experts anticipate Software Defined Storage (SDS) and hyperconverged infrastructure (HCI) will be significant areas for investment in 2016 and beyond as organizations seek to manage their rising storage and compute requirements, especially those hindered by a flat or declining IT budget.

In fact, a January 2016 survey from ActualTechMedia reported that 71 percent of enterprise organizations either are considering or are already deploying SDS or HCI to improve the agility of their data centers. IDC has predicted the hyperconverged market will pass the $2 billion mark this year and Gartner forecasts it will be worth $5 billion by 2019.

Many organizations are adding SDS tools to the capabilities of their existing storage devices to deliver modern data services, such as deduplication, stretched clustering and replication. Meanwhile, HCI eliminates the array altogether. Both SDS and HCI enable data center administrators to scale out by adding more nodes to expand capacity and performance, or to scale up by attaching more storage to existing nodes.

As IT leaders consider opportunities to implement SDS and HCI solutions, three main use cases emerge: remote office and branch office environments (ROBO); the introduction of new architectures into an organization; and virtual workspaces and desktops.

ROBO

ROBO environments face a number of challenges. One is configuration drift, where systems start to vary from the established baseline configuration for remote offices, leaving organizations struggling to support a variety of hardware and software configurations from site to site. In addition, costs and complexity associated with disaster recovery becomes an issue because ROBO locations are often located miles away from a central data center. They require a simple way to move and replicate data back to headquarters for business continuity. Coupled with the lack of IT personnel onsite, these challenges can put tremendous strain on organizations.

HCI can help organizations that have many ROBOs to deploy infrastructure in small, consistent blocks that can be easily and centrally managed. Some HCI appliances automatically replicate copies of data to centrally located SDS devices, which avoids extra hardware and bandwidth costs while improving application uptime.

New architecture

Hyperconverged appliances can also enable organizations to introduce new architecture and phase out older systems without disruption. For example, appliances can be used to save on costs and space in co-location scenarios while providing greater security, redundancy, reliability and savings. Enterprises have been able to double capacity for the same cost of adding disks to its legacy environment, boost performance, reduce its footprint and increase its compute power. These companies can also reuse their old equipment to create a disaster recovery site for data and virtual machines.

Taken a step further, for organizations reluctant to offload workloads to the public cloud because of security and privacy concerns, HCI can enable them to benefit from some of the advantages of cloud technologies and new application models on-premises. For some, it can be a way to bridge the gap to the public cloud.

Virtual Desktops and Workspaces 

In the case of virtualization and virtual desktop infrastructure (VDI), SDS can deliver significant benefits compared to traditional storage. As well as reducing the amount of storage hardware, it also supports larger numbers of users for less cost by using storage more effectively. Typically, SDS can be as much as 40-50 percent less expensive than traditional storage infrastructure, increase storage capacity by as much as 80 percent, improve the scalability of the solution and, in the case of VDI, deliver faster desktop login times.

SDS is also cheaper than using SAN and SSD to deliver VDI with better performance, more scalability, less rack space, less power and less technical support. If done effectively, organizations are able to deliver 200 IOPS per desktop on VDI, compared to 20 IOPS using SSD-based SAN storage and 70-80 IOPS for a physical desktop. In addition, it can also deliver much improved latency.

SDS and HCI bring significant levels of simplicity while turning the data center economic model on its head. Because data center tend to be hybrid or all-flash, they significantly outperform traditional disk-based storage arrays. HCI appliances integrate compute and storage into server nodes so they take up less space than traditional devices and because they deploy less equipment, power and cooling costs are reduced.

The ability to scale easily has liberated data administrators from buying three to five years’ worth of storage in advance and enabled adoption of a “just in time” approach. SDS AND HCI enable organizations to scale storage in small modular units as they grow their compute resources. For the first time, the advantages of using them together are giving IT leaders more reasons to put SDS and HCI on their project short list.

Ruben Spruijt, Field Chief Technology Officer at Atlantis Computing.

One Response to Software-Defined Storage and Hyperconverged Infrastructure: 3 Use Cases

  1. The main driver of HCI is simplicity. It is better suited for ROBO and VDI. Also HCI is a good fit for smaller deployments or a number of smaller deployments (thus good match for ROBO).

    SDS is way more powerful and cost efficient, but fits in larger or more complex use cases. You mention it: “some HCI appliances automatically replicate copies of data to centrally located SDS devices” – this is absolutely true. To look at proper SDS deployment you can either aim to replace a traditional SAN/all-flash array for the benefits – reliability, cost reduction and scalability and performance (is this what you meant by the second use case?). Or you can think of it as an upgrade of HCI when volumes raise and you need 10+ node deployments and it makes sense to segregate for easier management and cost optimization.

Add a Comment

EnterpriseAI