In order to enhance the independent innovation capability and satisfy a large number of basic research, Shenzhen City completed the Shenzhen Supercomputer Center in November 2011. The computer operation speed is 1271 teraflops per second, ranking second in the world. The total memory capacity of the system is 232 TB and the storage is 17.2 PB. This computing speed, the world's leading "super brain" has become an important engine for building a national innovative city in Shenzhen. Nowadays, various departments and enterprises are increasingly choosing to rent resources from the super-computing center. Therefore, the current resources can no longer meet the needs of the blowout, and it is urgent to build a new high-performance and high-reliability platform to improve the overall platform access computing power.
10 Gigabit backbone architecture needs to be upgraded
The existing network of the Shenzhen Supercomputing Center adopts a 10 Gigabit backbone architecture. With more and more resources, the current network architecture can no longer meet the existing needs. At the same time, with the adjustment and migration of resources, the existing network architecture maintenance costs are getting higher and higher, and the massive demand for cloud computing big data platforms is rather weak.
1. More and more resources are connected to the super-calculation center. How to deal with it?
With the resources of various functional departments in Shenzhen gradually switching to the platform of Shenzhen Supercomputing Center, the existing 10 Gigabit backbone network of the Supercomputing Center has not been able to meet the rapid growth of big data, and the network has become the bottleneck of data transmission.
2. What happens when resource migration leads to frequent network configuration adjustments and maintenance costs are rising?
As more and more resources are connected to the super-cloud platform, it will inevitably bring about the creation and release of resources. On the basis of the existing architecture, each resource adjustment needs to be adjusted accordingly. Coordination requires multi-departmental cooperation, and the human and material resources are expensive and the maintenance cost is high.
Huawei data center network, let the supercomputer center fly
In the super-computing cloud computing network system in Shenzhen, Huawei provides a high-bandwidth and high-reliability solution with 40GE cores, which can effectively solve the business scenarios of large-scale east-west traffic and concurrent access of hyper-large-scale virtual machines in cloud computing scenarios. The operation of the cloud computing platform of Shenzhen Supercomputer Center.
1. Core equipment is recommended for high performance design
The solution adopts the high-end data center switch CE12800 that supports the maximum switching capacity of 160Tbps, and the data center high-end firewall USG9560 with up to T-class processing performance. It supports 40GE full-speed forwarding to ensure non-blocking data transmission. The high-density 40GE/100GE boards are considered. The need for future business expansion.
2. Huawei virtualized TRILL technology to build a loop-free network
The TRILL technology of the CE series data center switches virtualizes all network devices into one fabric network. The command configuration is simple and easy to extend all physical links. While building a loop-free network through TRILL technology to ensure reliability, TRILL technology also ensures smooth migration of virtualized resources, which simplifies operation and maintenance and reduces costs.
3. Storage network convergence
The LAN/SAN converged technology simplifies the network, unified switching Ethernet high-speed forwarding, and takes full advantage of Ethernet performance and bandwidth. It reduces investment and simplifies the wiring structure of the data center.
4. Security and other value-added services are deployed in series to ensure security.
In the network solution, the data center firewalls are deployed in a series mode. The deployment mode meets the policy control of high security. The gateway is deployed on the firewall. By deploying virtualization technologies and using virtual firewalls, the network security of the super-computing center is effectively guaranteed.
In the Shenzhen Supercomputing Network System, Huawei provides a high-performance, high-reliability, strong integration, virtualization, easy-to-manage and one-stop overall network solution for the Shenzhen Supercomputing Center cloud computing platform. The entire network 10 Gigabit to convergence, 40GE to the core, greatly improved the operational efficiency of the Shenzhen supercomputer cloud computing platform.
With the increasing deployment scale of the application system of Shenzhen Supercomputing Center and the increasingly complex deployment model, Huawei's solution realizes the virtualization and integration of hardware resources, unified management and on-demand scheduling, and carries out efficient operation and maintenance and energy management. At the same time, to provide better support for the subsequent super-computing center services in Shenzhen functional departments. Huawei's data center network, let the "Shenzhen brain" fly!
Smd Speaker,Smd Subwoofer,Mini Smd Speaker,Smd Piezo Speaker
NINGBO SANCO ELECTRONICS CO., LTD. , https://www.sancobuzzer.com