D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM '93, pp. 722--731, March 1993.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Connectionless(CL) service is the predominant mode of communications in LANs MANs. However, ATM networks are inherently connectionoriented (CO) In order to support the CL data service on B ISDN, two alternatives, direct and indirect approaches for the ATM networks have been identified by CCITT[12 ]. For direct approach, CL service is provided directly within the B ISDN. Connectionless data traffics between the B ISDN customers are transferred by way of the connectionless servers (CLSs) in which connectionless service functions (CLSFs) are provided. Connectionless protocol functions such as ....

....have the benefit in bandwidth efficiency and scalability (due to the effect of statistical multiplexing) it suffers from much more delay. This is because of the more intermediate endpoints, the more the delay for data processing, routing (and segmentation, reassembly for the reassembly mode server[12]) at each CLS. In general, a smaller p will result in a higher multiplexing gain if taking more than one CLS is not prohibited. However, the smaller p implies a larger total transport cost. A good compromise can be found, in addition to introducing CLS cost under a certain budget consideration, by ....

D. F. Box, D. P. Hong and T. Suda, Architecture and Design of Connectionless Data Service for a Public ATM Network, IEEE INFOCOM, pp. 722-731, 1993.

....Ethernet or FDDI. Because this emulation is transparent to the network layer, LANE supports many different network protocols, in addition to IP. For the purposes of this research, there are few differences between an ATM network running LANE and one running IP over ATM. A third model, discussed in [Box93] and [Omundsen94] proposes connectionless servers within ATM networks to handle datagrams such as those generated by IP. In this model, end hosts establish virtual circuits to connectionless servers within the ATM network, which forward datagrams in much the same way as IP routers. Virtual ....

D. Box, D. Hong, and T. Suda. "Architecture and Design of Connectionless Data Service for a Public ATM Network", Proceedings of INFOCOM `93, San Francisco, CA, March 1993.

....it is more easily accommodated by an ATM network. In a public ATM setting where a large number of interconnected LANs will exist, the direct approach is the most appropriate. Previous work by the authors has focused on buffering techniques and a preliminary architecture for connectionless service [9]. This paper focuses specifically on the design, implementation and performance of connectionless servers for the direct approach. In section 2, issues involved in connectionless server design are discussed. In section 3, the two connectionless server forwarding modes described in section 2 are ....

D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM '93, pp. 722--731, March 1993.

....refer to the connections leading into and out of the CLS as ports . With N incoming and outgoing ports, the CLS functions as an N 2N packet switch. The transmission of datagrams through connectionless servers requires a framing architecture (e.g. Connectionless Interworking Protocol (CLIP) [7]) that absorbs the differences in packet format and size differences. Connectionless servers operate above and within the AAL layer receiving ATM cells and forwarding them based on network layer addresses, such as IP, NSAP, or E.164, of the destination ATM end systems. The cell headers are ....

.... the larger datagrams into several cells, placing the ATM end system address into the first cell of a packet (the beginning of message or BOM cell) and associating the cells together when the datagram is forwarded [8] It was shown previously that the CLIP header can fit in one cell payload [7]. An encapsulated IP header can fit in the first cell of a segmented datagram as well. Since not all cells contain the destination address, the cells belonging to the same datagram must follow the same path. Thus, the primary task of a connectionless server is to route datagrams based on the ....

D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM '93, pp. 722--731, March 1993.

No context found.

D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM 93, March, 1993.

....cell based forwarding servers. It also enables the connectionless server to drop an entire frame if a single cell is damaged, thereby reducing useless traffic in the network. However, it makes buffer size and buffer management more important concerns and imposes several types of delays on frames [8]: ffl Reception delay: caused by delaying the transmission of the first cell of a frame until the frame s last cell arrives at the connectionless server. This delay is at least as long as the transmission time of the frame. ffl Queueing delay: caused by delaying the transmission of a frame until ....

D. F. Box, D. P. Hong, and T. Suda, "Architecture and Design of Connectionless Data Service for a Public ATM Network," Proceedings of IEEE Infocom, vol. 2, pages 722--731, San Francisco, Calif., March 1993.

....the connections leading into and out of a CLS are referred to as ports . With N incoming and outgoing ports, a CLS functions as an N 2N packet switch. The transmission of datagrams through connectionless servers requires a framing architecture (e.g. Connectionless Interworking Protocol (CLIP) [4]) that absorbs packet format and size differences. Connectionless servers operate within and above the AAL layer receiving ATM cells and forwarding them based on network layer addresses, such as IP, NSAP, or E.164, of the destination ATM end systems. The cell headers are insufficient for this task ....

.... several cells, placing the ATM end system address into the first cell of a packet (the beginning of message or BOM cell) and associating the cells together when the datagram is forwarded [5] It was shown previously that the CLIP header and its encapsulated IP header can fit in one cell payload [4]. Since only the BOM cells contain the destination address, the cells belonging to the same datagram must follow the same path. Thus, the primary task of a connectionless server is to route datagrams based on the destination addresses contained in the BOM cells and ensure that the remaining cells ....

D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM '93, pp. 722--731, March 1993.

....cell of a packet and stored in a table as this cell is transmitted. Successive cells simply look up their next hop information in this table. This permits cells to be delivered before the entire packet has been received while ensuring that all the cells of the same packet follow the same path [7]. This is referred to as a streaming mode of operation. Two types of flow control are static and dynamic. Static flow control schemes do not react to congestion but rather attempt to prevent congestion from occuring in the first place. Examples include the leaky bucket and constant rate control. ....

D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM '93, pp. 722-- 731, March 1993.

No context found.

D.F. Box, D.P. Hong and T. Suda, "Architecture and design of connectionless data service for a public ATM network," In Proceedings of INFOCOM '93, pp. 722--731, March 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC