Shulou(Shulou.com)06/02 Report--

Super converged Infrastructure (HCI) has become very popular in the data center space, but it is reasonable that it is not suitable for every use case.

The biggest disadvantage of hyper-converged infrastructure is the lack of flexibility. This vendor-preconfigured system can provide consolidated server, storage, and network resources on a single node, but not all workloads are the same. Some applications require a lot of storage, while others consume a lot of network bandwidth. In addition, the super-converged infrastructure relies on virtualization so that multiple applications can share the same limited resources.

Therefore, when evaluating a super-converged infrastructure use case, IT professionals must pay close attention to their capacity planning and their scalability requirements, especially big data applications and Virtual Desktop Infrastructure (VDI).

Economics of HCI use cases

In the third quarter of 2017, global super-converged infrastructure revenue grew 68% year-on-year, with sales reaching $1 billion, according to IDC.

This IT infrastructure system is attractive because it integrates server, storage, and network hardware, and virtualization software, creating a preconfigured, easy-to-deploy package. IT departments can simply put these devices on their enterprise network and load their applications. Data center administrators spend very little time configuring these devices, which before that was a lengthy and time-consuming process.

"HCI enables companies to deploy new systems in the form of cloud computing." Industry analyst Stephanie Lang said.

However, the economics of some super-converged infrastructure use cases may not be attractive, especially if the workload is heavily dependent on a particular component. Big data is an area where super-converged infrastructure is not very suitable because these applications store a large amount of data and often require more storage space than the server's processing capacity.

Christian Perry, research manager at research firm 451, said: "most companies find that dynamic workloads such as Hadoop and (SAP) HANA do not perform very well on super-converged infrastructure."

VDI may have a similar problem because virtual desktops bring a high server CPU load. Although VDI was one of the largest superconverged infrastructure use cases in the early days, a survey by a consulting firm found that VDI is no longer one of the five most popular HCI workloads. Some companies have encountered problems with scale in the deployment of VDI. Compared to HCI, many companies have been trying to use GPU to handle VDI workloads.

If an organization's virtual desktops require more processing power, it is easier to scale with GPU than to purchase additional hyper-converged systems without unnecessary storage and network resource costs.

Hyper-converged infrastructure vendors have also been working to make their systems more responsive to fluctuations in infrastructure needs, trying to enable organizations to purchase a desired element without adding other components.

Limitations of Virtualization

The essence of a super-converged infrastructure is a virtualized system that enables organizations to maximize server capacity by running multiple applications on a single server, but may also cause problems with high-performance workloads.

For example, a company may want its e-commerce applications to work without interference. But if it runs on a virtual server without properly planned capacity, other applications may take resources away from it, causing it to slow down or crash. This is a potential trap for any virtualized deployment, not just for superconverged infrastructure use cases.

To avoid such problems, most companies may choose to run some applications on physical servers, in which case super-converged systems often cannot provide support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.