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

The infrastructure can be combined to treat compute, storage, and network devices as a pool of resources that are provided on demand according to the best performance requirements of different workloads. This is an emerging infrastructure designed to optimize IT resources and increase business agility.

This approach is similar to the public cloud in responding to requests from shared IT resources-except that the composable infrastructure is located in the enterprise's local data center.

IT resources are treated as services, and composability means to make them available dynamically according to the needs of different physical, virtual, and container applications. The management is designed to discover and access compute and storage pools to ensure that the right resources are in the right place at the right time.

Ric Lewis, Senior Vice President and General Manager of HPE Software definition and Cloud Computing Group, pointed out that the goal is to reduce underutilization and oversupply when creating a more flexible data center, and to provide a collaborative and composable infrastructure platform.

"when customers log in to the public cloud, they get a set of resources: computing, storage, networking. To put it simply, I need so many resources to run the application, please give it to me according to my requirements." Lewis said: "when my application is finished, I will give these resources back to you, and you can let others use them. What we do with composable infrastructure is to build it into the platform, and we can do the same dynamic resource sharing."

Combinable VS converged vs super converged infrastructure

A unit of the converged infrastructure contains a preconfigured software and hardware package that simplifies procurement and is easier to operate and maintain than traditional servers, storage, and network switches. Converged infrastructure is usually designed for specific applications or workloads, and when computing, storage, and network components are physically integrated, the management of these discrete resources is usually silo-based.

Superconvergence adds deeper abstraction and higher levels of automation, making infrastructure resources easier to use. In a super-integrated environment, software-defined elements can actually be implemented by seamlessly integrating into the hypervisor environment. Organizations can expand their capacity by deploying additional modules.

Like converged or hyper-converged infrastructure, combinable infrastructure combines computing, storage, and network structures into a single platform. However, it is not pre-configured for specific workloads like converged or hyper-converged infrastructure.

"if you want to implement software-defined storage in virtualization, hyperconvergence can solve the problem very well," Lewis said. However, for data center customers, they are not only doing a virtualized environment, but they are not doing everything on software-defined storage. What they are doing is on a large scale, running virtual machines and bare metal. Customers want to provide a simple environment for VM, bare metal, containers, and new cloud-native applications. "

Superconverged infrastructure also has scalability limitations, and a typical superconverged environment can scale to 20 or 30 nodes. "Super convergence is good, but it doesn't solve all workloads, and it doesn't scale to the level that customers want to scale," Lewis said. "

The combinable infrastructure allows for further scalability and allows enterprises to have more pools of mobile resources.

Flexible allocation of resources

In a composable infrastructure, resources can be reconfigured to form an infrastructure environment of the exact size required by each workload. For example, developers can request virtual machines through any combination of compute, network, and storage capacity, and when the workload ends, these infrastructure resources are sent back to the pool for other users to access. One workload may be a compute-intensive application that requires a lot of CPU power, while another workload may require more memory.

"the application can get whatever it needs while it is running, and when it is finished, these resources will be returned to the pool. It is not just focused on running VM like a super-converged environment." Lewis said. As a result, the local infrastructure looks more like a public cloud IaaS environment.

HPE was one of the first vendors to provide composable infrastructure platforms.

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.