Cloud Trends for High-performance Computing

High Performance Computing Overview high-performance Computing (HPC) is the computing system and environment that typically uses many processors (as part of a single machine) or several computers that are organized as a single computing resource in a set of clusters. There are many types of HPC systems ranging from large clusters of standard computers to highly specialized hardware. Most cluster-based HPC systems use High-performance network interconnects, such as those from InfiniBand or myrinet. The basic network topology and organization can use a simple bus topology, in a very high performance environment, mesh network systems provide a shorter latency between the host, so you can improve overall network performance and transmission rate. This is about the content of High-performance computing, the advent of high-performance computers has been over 70 years, the speed of the computer and computing power has increased 1 million times times.

Narrow cloud computing refers to the delivery and usage patterns of IT infrastructure, which means acquiring the required resources through the network on an as-needed and extensible basis; broad cloud computing refers to the delivery and usage patterns of services, which means that the required services are obtained through the network in an on-demand, extensible manner. This service can be it and software, Internet-related, but also other services. The core idea of cloud computing is to manage and dispatch a large number of computing resources with network connection, and form a computing resource pool to serve the users on demand. A network that provides resources is called a "cloud." The resources in the cloud can be extended indefinitely by the user, and can be obtained at any time, on demand, at any time, and paid for by use. Cloud computing industry level three tier: Cloud software, cloud platform, cloud equipment. Now cloud computing is no longer a novelty in the industry, and many technologies are seeking to combine with the cloud, and High-performance computing is no exception.

The current HPC cloud implementation, considering the role of the following aspects, customers can pay according to the running time of the application or the time to use storage and data transfer services. Infrastructure is almost instantly available, and conversely, the infrastructure can shrink back to nearly 0 resources when there is no workload. The flexibility of resource allocation can be dramatically improved by assigning the type of operating system or server at any time, depending on the workload.

But there are still some problems in this process, said Guojie, academician of the Chinese Academy of Engineering, "the ease of use of cloud computing will affect the traditional HPC computing model, and on-demand cloud computing provides a more interactive computing model for HPC." The key technology used in cloud computing is virtualization, which allows resources to be scheduled on demand, but also leads to a loss of speed. While High-performance computing is the pursuit of "high-speed", the two seemingly contradictory. It is expected that the flexible application model of cloud computing can solve the problem of system management of public high-performance computing platform.

Implementing a HPC cloud requires several tools, including the virtual machine hypervisor platform, the workload manager, and the Infrastructure Management Toolkit. The Management toolkit should provide functionality such as policy definition and execution, configuration management, resource reservation, and reporting. The virtual machine management platform should provide a good foundation for the virtual parts of cloud computing resources. Finally, the workload manager should provide job management functionality.

The future of cloud computing with high-performance computers will certainly be the road to a more smooth, there are more service patterns, in addition to the choice of High-performance computing cloud solutions should be evaluated in terms of rationality, security, etc., only such a high-performance calculation of the cloud trend will become a convention.