Network bandwidth control

Details Network bandwidth control Network bandwidth control

1997-12-22

2001-06-19

Patel, Ajit (Department: 2661)

Multiplex communications

Communication techniques for information carried in plural...

Adaptive

C370S231000

Reexamination Certificate

active

06249530

ABSTRACT:

BACKGROUND OF THE INVENTION
This invention relates to bandwidth control over a telecommunications network. In particular, the invention relates to bandwidth control for controlling communication flow on a connection to a network station, for example for controlling TCP/IP traffic on an Internet connection.
Conceptually, the Internet provides three sets of services. At the lowest level, a connectionless delivery system provides a foundation on which everything rests. At the next level, a reliable transport service provides a high level platform. At the third level, application services are provided which rely on the reliable transport service.
A fundamental Internet service consists of an unreliable, best-effort, connectionless, packet delivery system. The service is described as being “unreliable” because delivery is not guaranteed. A packet may be lost, duplicated, or delivered out of order, but the Internet will not detect such conditions, nor will it inform the sender or receiver. The service is described as being “connectionless” because each packet is treated independently from all others. A sequence of packets sent from one machine to another may travel over different paths, or some may be lost while others are delivered. The service may be described as “best-effort” because the Internet aims to deliver packets.
The protocol that defines the unreliable, connectionless, delivery mechanism is called the “Internet Protocol”, and is usually referred to by its initials IP. IP defines the formal specification of data formats, including a basic unit of data transfer and the exact format of all data passing across the Internet. IP also includes rules which specify how packets should be processed and how errors should be handled. In particular, IP embodies the idea of unreliable delivery and packet routing.
Above the IP layer of the Internet protocol structure one service which is provided is a reliable transport service which is typically called the “reliable stream transport service”, defined by the Transmission Control Protocol (TCP). The combination of the TCP protocol and the underlying Internet protocol (IP) is often referred to as TCP/IP.
The reliable stream delivery service provided by the TCP can be contrasted with the unreliable datagram protocol (UDP) which is also provided over the Internet. The UDP provides an unreliable delivery service because delivery is not guaranteed. For example, packets may be lost or destroyed when transmission errors interfere with data, when network hardware fails, or when networks become too heavily loaded to accommodate the load presented.
The TCP on the other hand has a complex structure providing delivery by means of a stream of bits, divided into eight-bit bytes. The TCP specifies the format of the data and acknowledgements that two computers are to exchange to achieve reliable transfer, as well as the procedure to ensure that data arrives correctly.
Further details of aspects of the Internet and TCP/IP protocols may be found, for example, in the following U.S. Pat. Nos. 5,293,379; 5,307,347; 5,307,413; 5,309,437; 5,351,237; and 5,535,199.
As mentioned above, given that the underlying Internet protocol is unreliable, TCP transmissions operate in accordance with a technique known as positive acknowledgement with retransmission. The technique requires a recipient to communicate with the source, sending back an acknowledgement message every time it receives data. The sender keeps a record of each packet that it sends and waits for an acknowledgement before sending the next packet. The sender also starts a timer when it sends its packet and retransmits a packet if the timer expires before the acknowledgement arrives.
The period between the transmission of a message and the receipt of an acknowledgement (say A
1
) is termed the Round-Trip-Time (RTT). The RTT varies over time depending upon many factors such as, for example, network loading (e.g., delays at intermediate nodes in the system) and loading on the receiver. An important factor in determining the RTT is the available bandwidth. Thus, where multiple clients have access to a common server, for example, in order to balance the quality of service between clients, it is desirable to control the bandwidth allocated to the individual clients.
A bandwidth control mechanism has been proposed which is based on the measurement of the RTT experienced on a given connection and the advertised TCP window size of the receiver which is to be found in the WINDOW field of a TCP header. The RTT can be defined as the time between sending a byte with a particular sequence number and receiving an acknowledgment that covers that sequence number.
With such an arrangement, the maximum amount of bandwidth that has been used by a particular connection since an arbitrary time t(i−1) depends on the value of both parameters according to the expression known as the “Bandwidth-Delay Product”:
Used Bandwidth since t(i−1) bps=Receiver Window Size since t(i−1)/ RTT(S)  (1)
This expression presents a direct way of modifying the bandwidth usage of a particular flow. In other words, if we know the RTT of the flow, we can act on the Window Size value to get the desired bandwidth.
However, the computation described above has limitations. Firstly, there is the problem of determining the value of the RTT with sufficient precision. The computation of the RTT is not an easy task as it requires arrival of packets in both directions between two stations to be taken into account. Also, the RTT is not the same for every packet (e.g., a TCP acknowledgment may acknowledge the reception of various packets). Moreover, an average (non-instantaneous) calculation must be adopted. Further information on the so-called bandwidth delay product can be found in, for example, TCP/IP Illustrated, Volume 1: The Protocols W. Richard Stevens Addison-Wesley, October 1995.
Although TCP is provided over the Internet, it is in fact an independent general purpose protocol which can also be used with other delivery systems. TCP makes very few assumptions regarding the underlying network, and it can also be used over a single network like Ethernet, as well as over a complex Internet, or Intranet. Indeed, although particular implementations of the invention are described in the context of the TCP/IP traffic over the Internet, the invention is not limited thereto.
Accordingly the present invention seeks to provide a solution to bandwidth control without the disadvantages of the approach described above.
SUMMARY OF THE INVENTION
Particular and preferred aspects of the invention are set out in the accompanying independent and dependent claims. Combinations of features from the dependent claims may be combined with features of the independent claims as appropriate and not merely as explicitly set out in the claims.
In accordance with an aspect of the invention, there is provided a computer implemented method of controlling network bandwidth by dynamically determining a current window size for a destination buffer as a bandwidth control parameter for a communication flow. The method comprises:
a) determining a first window size for the destination buffer at a first time; and
b) determining a current window size at a second time as a function of the first window size, a desired bandwidth value and a bandwidth used between the first time and the second time.
Thus, a method according to the invention provides a solution to the problem of achieving a desired bandwidth for a particular connection, without the problems associated with RTT calculations as described in the introduction. An embodiment of the invention enables a “desired bandwidth” limit to be assigned to a specific connection, for example a TCP connection, and for the adaptive determination of a window size for a communication flow without needing to perform absolute RTT calculations. In other words, whereas the approach described in the introduction involves the control of bandwidth in response to calculation of an absolute RTT value, an embodiment of the present i

Affiliated with

Also associated with

Rating

Network bandwidth control does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Network bandwidth control, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Network bandwidth control will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2458514