State of the art networking

1935 Words4 Pages

The growth of network bandwidths has contributed to Internet traffic since the introduction of packet switching. In recent years, new applications such as Peerto- Peer (P2P) file sharing, multi-media, and mobile computing have increased users’ expectations, motivating new designs in which various communication links, such as grid [32], satellite, wireless and mobile computing, can securely participate and handle traffic at higher layers of the protocol stack. These new applications vary in traffic and connection characteristics in various communication links. Most of them still use TCP for data transfer because of its reliability and stability. There have also been performance issues in the implementation of large networks that require high bandwidths. These issues led to the development of new and different schemes with more reliable characteristics and better congestion control to address the problems; one example is XCP (i.e. Xplicit Congestion Protocol) [23]. XCP demonstrates good performance characteristics when tested on routers and on satellite systems [23,24]. Other variants, such as STCP and DCCP, are designed to improve congestion control. Some variants such as XCP face challenges on deployment because they require changes in the routers in addition to the operating systems of end hosts. Recent studies have shown that the gradual deployment to update Internet facing routers results in a significant performance drop. With characteristics similar to TCP, it also has security flaws. Apart from congestion control and performance, for which TCP variants were developed, security considerations need to be included in the architectural designs of the new generation of protocols [23]. Developments in 2007 i... ... middle of paper ... ...ent from that of traditional Internet applications. First, the data transfer often lasts a very long time at very high speeds. Second, distributed applications need cooperation among multiple data connections. Fairness between flows with different start times and network delays is desirable. Finally, in grid computing over high performance networks, the abundant optical bandwidth is usually shared by only a small number of bulk sources. The concurrency is much smaller than that on the Internet. Here adopt an example presented by []. We present a simple but typical example application – called the streaming join. Suppose that real time data streams coming from a remote machine A and a local machine B are joined at another local machine C with a window-based join algorithm [71]. Also, assume that the two data streams are composed of records of the same size.

More about State of the art networking

Open Document