Transcript View/Open
A.SATHEESH Department of Software Engineering Periyar Maniammai University Tamil Nadu SUMMARY What is Data Center ? Switch architecture in Data Center . Network switch. Buffer management in network switch . Static Memory Management in Data Center Dynamic Memory Management in Data Center Simulation Model System Requirements Result and Discussion Conclusion References DATA CENTER A data center is a facility used to house computer systems and associated components, such as telecommunications and storage systems. It generally includes redundant or backup power supplies, redundant data communications connections, environmental controls (e.g., air conditioning, fire suppression) and security devices. Network Switch A network switch is a computer networking device that links network segments or network devices. The term commonly refers to a multiport network bridge that processes and routes data at the data link layer of the OSI model. Buffer in Network Switch Static Memory Management in Data Center It has a central packet processing and classification engine. Each port is connected with individual static buffer. Disadvantage In static buffer once the buffer is full, additional in coming packets will be dropped Large buffers in network switches can considerably add to the system cost, operational complexity, resulting in less deterministic and impulsive application performance. Dynamic Memory Management in Data Center It has a centralized, fully sharable, dynamically allocated and adaptive buffer. Advantages It deliver s most favorable buffer utilization and burst absorption for data center workloads. cost effective performance, which is suitable for modern data center switches in cloud applications. System Requirements 64 x 10 Giga bit Ethernet switch device Fully Meshed Topology configuration Packets size is subjected from 64 to 1500 bytes Minimum Threshold and Maximum threshold value are set as 20% and 80% queue buffer size. The 64 x10 port is replaced with 96x10 Giga bit Ethernet and 32 x 40Gigabit Ethernet switch port configurations and tested with the previous conditions. Uniform random distribution across all ports with, a loading factor of 80 % on all ports, and a target frame loss rate of 0.1% is taken Simulation Model STATIC AND DYNAMIC BUFFER UTILIZATION PORT SIZE GbE SIZE OF DYNAMIC BUFFER UTILIZATION USAGE MB SIZE OF STATIC BUFFER UTILIZATION MB 64 X 10 5.5 25.9 96 x 10 8.55 37.8 32 x 40 3 15.8 Frame Loss with Static and Dynamic buffer Frame loss occurs when one or more frame of data travelling across a computer network fail to reach their destination. This frame loss can be caused by a number of factors including signal degradation over the network medium due to multi-path fading, packet drop because of channel congestion , corrupted packets rejected in-transit, faulty networking hardware, faulty network drivers or normal routing routines Burst Absorption with static and dynamic buffer Burst size is an amount of data that may be transferred without being affected by the other limitations Conclusion A switch using static buffer allocation require up to five times as much packet memory as a switch using dynamic shared buffer pool . While using dynamic buffer the burst size obtained is 3 to 6 times greater than the burst size obtained using static buffer . REFERENCES [1 ] M. Al-Fares, A. Loukissas, and A. Vahdat. A scalable, commodity data center network architecture. In SIGCOMM, 2008. [2] A. Greenberg et al. VL2: A scalable and flexible data center network. In SIGCOMM, 2009. [3] C. Guo et al. Bcube: High performance, server-centric network architecture for data centers. In SIGCOMM, 2009. [4] J. Hamilton. On designing and deploying Internet-scale services. In USENIX LISA, 2007. [5] R. Morris, “TCP behaviour with many flows,” IEEE International Conference on Network Protocols (ICNP’97) [6] R. Morris, “Scalable TCP congestion control,” IEEE Infocom’00, Tel Aviv, Israel, Mar. 2000. [7] Al-Fares, M., Radhakrishnan, S., Raghavan, B., Huang, N., Vahdat,A.: Hedera: dynamic flow scheduling for data center networks. In: Proceedings of the 7th USENIX Symposium on Networked Systems Design and Implementation (NSDI ’10), San Jose, CA, April 2010 [8] IEEE 802.1 Data Center Bridging Task Group, available at: http:// www.ieee802.org/1/pages/dcbridges.html (2011) THANK YOU