W R. Stevens. TCP/IP Illustrated Volume 1. Addison Wesley, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....violations. Techniques based on LIMD have been used in many systems to adapt to changing system conditions. To illustrate, TCP employs a congestion control algorithm based on LIMD; the LIMD algorithm is used to compute the congestion window size based on the bandwidth available for the connection [15]. The window size in increased in a linear fashion so long as bandwidth is available and no loss occurs. The window size is reduced by half (a multiplicative decrease) when a loss event occurs (indicating the presence of congestion) Such an algorithm allows TCP to probe the network for the ....

W R. Stevens. TCP/IP Illustrated Volume 1. Addison Wesley, 1994.

....section, we briefly describe the working of a typical HTTP server on a machine with a UNIXbased TCP IP implementation. The description provides background for the discussion in the following sections. For simplicity, we focus our discussion on a BSD based network subsystem [McKusick et al. 1996; Wright and Stevens 1995]. The behavior of other implementations of TCP IP, such as those found in UNIX System V [Bach 1986] and Windows NT [Custer 1993] is similar. In the HTTP protocol, for each URL fetched, a browser establishes a new TCP connection to the appropriate server, sends a request on this connection and ....

Wright, G. and W. Stevens (1995), TCP/IP Illustrated Volume 2 , Addison-Wesley, Reading, MA.

....to denote either client processes or client threads, as this distinction makes no difference to our method. 6 Banga and Druschel, Measuring the Capacity of a Web Server under Realistic Loads Bestavros 1996] and with peak rates exceeding the average rate by factors of 8 to 10 [Mogul 1995a; Stevens 1996]. Furthermore, peak request rates can easily exceed the capacity of the server. By contrast, in the simple request generation method, a small number of clients have independent think time distributions with small mean and variance. As a result, the generated traffic has little burstiness. The ....

Stevens, W. (1996), TCP/IP Illustrated Volume 3 , Addison-Wesley, Reading, MA.

....We are not the first to point out the performance implications of the the interaction between the Nagle algorithm and the delayed acknowledgment mechanism. The effect occurs in contexts besides NNTP. Early examples including the use of multi byte function keys in interactive terminal sessions [18], and the interactive X window system [17] In both of these cases, implementors quickly realized the need to disable the Nagle algorithm [3] More recently, the introduction in HTTP of persistent connections (the use of a single TCP connections for several HTTP requests) has led several ....

W.R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, Reading, MA, 1994.

....control blocks are timed out after a configurable period of inactivity following the splice or on receipt of a cleanup message from the port controllers. Besides the functions described above, the CPU has to deal with a few subtle problems while splicing TCP connections. Handling of TCP options [15] is one of the issues that deserves special attention. TCP supports various optional functions such as maximum segment size (MSS) window scale factor, timestamp, and selective acknowledgment (SACK) 10] These options are negotiated during the initial TCP handshake and the set of options ....

W. R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, New York, 1996.

.... in the acknowledgement packets, which bounds the number of unacknowledged bytes in transit from the sender to the receiver, and (d) congestion control is achieved by a combination of four mechanisms: round trip time estimation, slow start, congestion avoidance, and fast retransmit fast recovery [8]. TCP is inherently unicast in nature and cannot be used for reliable multicast communication without special network support for a variety of reasons: ffl The 3 way handshake is inherently unicast, involving the exchange of initial sequence numbers (ISNs) and maximum segment size (MSS) Since ....

....P120 TI Travelmate 192.17.9.2 192.17.8.1 192.17.7.1 192.17.9.1 link a 192.17.8.2 link b fast receiver slow receiver 224.1.1.12 240.1.1.12 Fig. 4. Testbed configuration Filter (BPF) and the packet capture library (libpcap) that is built upon it. Most BSD derived kernels support BPF [8], which allows a portability between different platforms. Since BPF forwards a copy of the filtered packet rather than the packet itself, we use the IP firewall to prevent ACKs from getting forwarded by default. IP firewall rules contain the packet type to be filtered and one of three actions for ....

W. R. Stevens, TCP/IP Illustrated Volume 1, Addison Wesley, March 1996.

....it is clear that we can use the port number for this. Since the port numbers are stored as 16 bit integers the port number 20 results in the hexadecimal byte pattern 0014. The second thing to be known is the position for the match within the AAL frame. A look at the TCP IP header specification [Stevens, 1994] shows that the TCP header starts 20 bytes after the beginning of a IP packet and that the source port occupies the first two bytes, the destination port the second two bytes within a TCP header. Therefore as shown in Figure 3.4 the right position for starting the match is an offset of 20 bytes ....

Stevens, W. R. (1994). TCP/IP Illustrated Volume1 (The Protocols), volume 1. Addison-Wesley.

....start is by looking at the network protocol layers and how they are affected. The discussion that follows assumes a basic knowledge of computer networking and layered network protocols. Any reader not familiar with the diagram in figure 1. 1 should review these principles in the following sources [50 52, 55]. 1.3 Protocol layers involved The architecture of a wireless network requires the careful integration of the various pieces into one unified and transparent system. Ideally, a change in the physical medium would only require changing the physical layer. Limiting changes to the physical layer ....

....on protocol specifications, referred to as an RFC, which will handle these problems [43] Network layer concerns are not addressed further in this document. However, the reader is encourage to consult the following 11 sources for a better understanding of the issues and concerns of this layer [6, 21, 26, 29, 30 33, 37, 42, 43, 49, 50 52, 54, 55]. 1.4 Desired traits or features in a wireless local area network (WLAN) A look at the desired features in a wireless local area network will provide further insight into the architectural issues of this form of networking. Some of the features that would be beneficial in a wireless network can ....

Stevens W. R., TCP/IP Illustrated Volume 1. (Addison-Wesley Publishing Co.: Reading, MA, 1994).

....second packet to queue until the first packet is transmitted on the slowest link in the transmission path. Therefore, it is proportional to the bottleneck bandwidth. Several monitoring tools use this technique [5,6,7,11] For example, Pathchar [7] uses this method in combination with traceroute [3], in order to derive per hop information about the network. The framework presented in [5] consists of two parts. The first part, b probe, similarly to the other tools above, only provides information about the bottleneck bandwidth using packet pair. The second part, c probe, provides an estimate ....

W. R. Stevens, TCP/IP Illustrated Volume 1 Addison-Wesley, Reading MA, Nov 1994.

....the TCP sender can now send that many more bytes of data. Delayed acknowledgments are particularly useful in client server applications, where data frequently flows in both directions. Many TCP implementations are derived from the 4. x BSD code base (described in great detail by Wright and Stevens [19]) These TCP stacks mark a connection that has just received new data by setting the TF DELACK flag, indicating that the connection needs a delayed acknowledgment. Subsequent transmission of an acknowledgment on the connection clears this flag. A separate timer driven background activity polls the ....

G.R. Wright and W.R. Stevens. TCP/IP Illustrated Volume 2. Addison-Wesley, Reading, MA, 1995. 7

....We are not the first to point out the performance implications of the the interaction between the Nagle algorithm and the delayed acknowledgment mechanism. The effect occurs in contexts besides NNTP. Early examples including the use of multi byte function keys in interactive terminal sessions [18], and the interactive X window system [17] In both of these cases, implementors quickly realized the need to disable the Nagle algorithm [3] More recently, the introduction in HTTP of persistent connections (the use of a single TCP connections for several HTTP requests) has led several ....

W.R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, Reading, MA, 1994.

....aggregate ACK upstream bypassing the TCP protocol stack. Capturing incoming SYN, FIN, RST and ACK packets belonging to a reliable multicast session is done by using the BSD Packet Filter (BPF) 7] and the packet capture library (libpcap) that is built upon it. Most BSD derived kernels support BPF [8], which allows a portability between different platforms. Since BPF forwards a copy of the filtered packet rather than the packet itself, we use the IP firewall to prevent original ACKs from getting forwarded by default. We deny (ignore the packet without sending ICMP message) incoming ACKs, RSTs, ....

W. R. Stevens. TCP/IP Illustrated Volume 1. Addison Wesley, March 1996.

....with a thoroughly checked and well studied implementation of TCP. To fully appreciate this advantage of using ns 2 for our experiments, we first review those features of TCP which are of particular interest for the purpose of this study. For a more comprehensive treatment of TCP, see for example [13]. TCP is one of the predominant transport layer protocols in the Internet, providing a connection oriented, reliable, byte stream between a source (or sender) and a destination (or receiver) TCP provides reliability by splitting the data into numbered segments. To insure that each segment is ....

W.R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, 1994.

....a window based flow control, TCP Vegas uses a rate based mechanism, which essentially estimates the available bandwidth per connection and sends packets with the corresponding inter packet interval. Most Internet applications nowadays use TCP for reliable transmission with delayed acknowledgement [109], where acknowledgements are sent every other packet. 3 Domain identifications are usually represented by the high bits of IP addresses. 22 Real time Multimedia Congestion Control Several congestion control designs have been proposed for real time data transmission, as compared to TCP for ....

W.R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, 1994

....A subnet is a logical arrangement of the IP addresses used. As UCCS uses a type B class IP address of 128.198.x.y, it makes it easy to use x as a subnet and y as a host on the subnet. Routers binds together the segments so that packets can traverse the entire network. According to Stevens[5] the term gateway should be used when talking about an application gateway that connects two different protocol suites like TCP IP and for instance IBM s SNA. The term router should be used when forwarding packets between segments. We will thus mainly be using the term router, but many of the Unix ....

....values are more or less taken out of thin air, except the TCP time out value. The TCP time out is determined by the TCP transaction specification which states that an endpoint should remain open in the state called TIME WAIT for the time of 2 MSL, where MSL is a system constant. According to [5], MSL is supposed to be 2 minutes, but common values are 30 seconds, 1 minute or 2 minutes. Linux uses 1 minute. Actually most values of the time out would have worked if the two hosts have a reliable and fast connection, where the incoming packet arrives before the rule times out. If the rule ....

[Article contains additional citation context not shown here]

[Stev94] W. Richard Stevens, TCP/IP Illustrated Volume 1, The protocols, ISBN 0-201-633469, 1994

....copies data between the connections until they are spliced, and finally splices the connection by sending the appropriate control messages to the port controllers. There are a few subtle issues that need to be dealt with when splicing TCP connections. One of them, viz the handling of TCP options [16] deserves special attention. TCP supports various optional functions such as maximum segment size (MSS) window scale factor, timestamp, and selective acknowledgment (SACK) 12] These options are negotiated during the initial TCP handshake and the set of options supported by both the client and ....

W. R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, New York, 1996.

.... occurrence of this highly complex scaling behavior of measured Internet traffic over small time scales by reducing it to and pointing out a plausible explanation in terms of previously observed phenomena in the dynamics of TCP type congestion control, among them ACK compression; see for example [12, 27, 26, 32, 17], or the more recent study [20] This empirical observation begs for a simple mathematical construction that incorporates the essence of flow control phenomena and leads to multifractal scaling behavior. Unfortunately, we have not yet succeeded in this endeavor and at this stage, referring to the ....

W.R. Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, 1994.

....when many clients are accessing it. Solutions using hierarchies of distributed directories can then be used, they can then cache results 26CHAPTER 2. PROPERTIES OF STRUCTURES FOR SERVERS of earlier requests to improve performance. This is a schema similar to the well known internet DNS service [Ste94]. Another, simplier distribution strategy is to store one field of a record for all records in a file entirely in one node and other fields on other nodes. However, if the amount of data grows fast, this is no scalable solution. We notice that these distributed data structures will also have to be ....

W. Richard Stevens. TCP/IP Illustrated Volume 1. Addison-Wesley, 1994.

....Simplified example of how select( is used of the kernel, and then scanned by the application. These costs clearly increase with the number of descriptors. Other aspects of the select( implementation also scale poorly. Wright and Stevens provide a detailed discussion of the 4. 4BSD implementation[23]; we limit ourselves to a sketch. In the traditional implementation, select( starts by checking, for each descriptor present in the input bitmaps, whether that descriptor is already available for I O. If none are available, then select( blocks. Later, when a protocol processing (or file system) ....

G. Wright and W. Stevens. TCP/IP Illustrated Volume 2. Addison-Wesley, Reading, MA, 1995.

....of protocol data units (PDUs) Note that LRP does not depend on a specific network adaptor or on an ATM networks. It can be used with any network and NI. The LRP architecture was implemented as follows. We modified the TCP UDP IP network subsystem that comes with the 4. 4 BSD Lite distribution [12] to optionally implement LRP. The resulting code was then downloaded into the SunOS kernel as a loadable kernel module and attached to the socket layer as a new protocol family (PF LRP) A custom device driver was developed for the FORE network adaptor. The 4.4 BSD Lite networking subsystem was ....

G. Wright and W. Stevens. TCP/IP Illustrated Volume 2. Addison-Wesley, Reading, MA, 1995.

....control is achieved by a combination of four mechanisms: round trip time estimation, slow start at the start of data transmission (or after a timeout) congestion avoidance near the stability point, and fast retransmit fast recovery. For details on TCP s operation, we refer the reader to [9]. The congestion control, flow control, and local recovery mechanisms of TCP can be naturally extended to support multicast communication by performing the cumulative aggregation of ACKs over the multicast tree rather than a single path. Additionally local recovery can be used to improve loss ....

....connection is done by using the BSD Packet Filter (BPF) 15] and the packet capture library (libpcap) that is built upon it. BPF allows a user level process to tap one of its network interfaces and capture any packets that match a user defined filter. Since most BSD derived kernels support BPF [9], it allows a portability between different platforms. 3. Blocking ACKs, FINs, and RSTs: Since BPF forwards a copy of the filtered packet rather than the packet itself, we use the IP firewall [16] to prevent ACKs from getting forwarded by default. IP firewall rules contain the packet type to be ....

W. R. Stevens. TCP/IP Illustrated Volume 1. Addison Wesley, March 1996.

....to protocol performance because the application is optimized for the network and vice versa. SRM suggests the notion of selective reliability which allows receivers to selectively repair portions of the data stream. However, there are problems in realizing this concept with traditional primitives [9] and they are addressed in a recent paper by Raman and McCanne [8] They propose SNAP, a scalable naming and announcement protocol, that provides a framework for hierarchical data naming. SNAP maintains a common syntax between the application and the transport which enables receiver tailored ....

Stevens, W. TCP/IP Illustrated Volume 1. Addison-Wesley, 1994.

....scanned by the kernel, subsetted, copied out of the kernel, and then scanned by the application. These costs clearly scale with the number of descriptors. Other aspects of the select( implementation also scale poorly. Wright and Stevens provide a detailed discussion of the 4. 4BSD implementation[23]; we limit ourselves to a sketch. In the traditional implementation, select( starts by checking, for each descriptor present in the input bitmaps, whether that descriptor is already available for I O. If none are available, then select( blocks. Later, when a protocol processing (or file system) ....

G. Wright and W. Stevens. TCP/IP Illustrated Volume 2. Addison-Wesley, Reading, MA, 1995.

No context found.

Richard, W.Stevens. (1994) TCP/IP Illustrated Volume 1. ISBN 0-201-63354-X, pp. 325 - 330; pp. 263 - 285, Addison-Wesley Professional Computer Series.

No context found.

[Wrig95] Gary R. Wright, W. Richard Stevens, TCP/IP Illustrated Volume 2, The Implementation, ISBN 0-201-63346-9, 1995

First 50 documents

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