Multiplex communications – Pathfinding or routing – Switching a message which includes an address header
Reexamination Certificate
1997-08-07
2001-09-25
Kizou, Hassan (Department: 2662)
Multiplex communications
Pathfinding or routing
Switching a message which includes an address header
C370S234000, C370S236000, C709S238000
Reexamination Certificate
active
06295294
ABSTRACT:
FIELD OF THE INVENTION
The invention relates generally to regulation of network traffic. More particularly, the invention relates to limiting congestion in packet switched network traffic such as that carried over the Internet.
BACKGROUND OF THE INVENTION
The Internet is becoming increasingly popular as more and more participants endeavor to benefit from the promised benefits of the Information Highway. However, just as the promise of unrestricted travel over the interstate highway system has been thwarted by traffic jams in many urban areas, congestion over the Internet backbone threatens to prevent deployment of new services and limit the enjoyment of current ones.
The Internet interconnects over one hundred thousand different networks into a vast global network. Already huge, the Internet is growing at an exponential rate. In fact, the Internet is growing so fast that some notable computing pioneers, such as Bob Metcalfe, the inventor of Ethernet, have predicted its collapse. Although the Internet has not collapsed so far, it is heavily congested and providing levels of service that are universally perceived as inadequate.
The problem with decreasing performance on the Internet has been addressed, at least in part, by changes in the speeds at which users interface with the Internet. An increasing number of users rely on high speed digital connections or high speed modems that communicate over the telephone lines at speeds thought not possible a few years ago. Unfortunately, the cumulative effect of higher speed connections is more data on the Internet and the problem of congestion is moved to the backbone of the Internet. In fact, many users disparagingly refer to the World Wide Web as the “World Wide Wait” because of the delays caused by, at least in part, increasing congestion.
Some users have sought to resolve these issues by changing their Internet Service Providers (ISPs) in an attempt to find the ISP with the highest perceived quality of service. However, the current switching technologies used for the Internet do not provide a guaranteed quality of service. Moreover, few, if any, ISPs are willing to guarantee a particular level of quality of service. Furthermore, except for the differences between a direct digital connection (e.g., ISDN) and an analog (modem) connection, there is no apparent direct relationship between the cost for the connection and the quality of service. Pricing differentials in contemporary ISPs depend primarily on whether a user selects a fixed rate with unlimited access (i.e., a flat monthly charge) or a rate based only on connection time.
Although consumers generally prefer flat-rate pricing, congestion on the Internet will likely be significantly worsened by flat-rate pricing. Because Internet data transfers can take place without human attention, there is little incentive against remaining connected and transferring large amounts of data that might or might not be useful. In the case of usage-sensitive pricing, however, the cost for each packet (or byte of data) precludes thoughtless bandwidth consumption. Accordingly, usage-based pricing provides a mechanism by which traffic on the Internet may be regulated. It also provides economic incentives for Internet service providers to build adequate capacity, incentives that are, at best, weak with flat-rate prices.
There have been several proposals for various kinds of usage-based pricing for the Internet, many of which are discussed in
Internet Economics
, McKnight, L. W., and Bailey, J. P., eds., (MIT Press, 1997), incorporated by reference herein. For example, MacKie-Mason and Varian have proposed imposing charges on packets that contribute to congestion in the network in “Pricing the Internet” in
Public Access to the Internet
, D. Kahin and J. Keller, eds., MIT press, 1995, pp. 269-314 and in “Pricing Congestible Network Resources,” IEEE J. Selected Areas Comm., 13 (1995), pp. 1141-1149. However, their proposal suffers from the disadvantage that it requires complex systems to conduct an auction among individual packets at the congested node. Moreover, their proposal does not address the absence of a direct correlation between delay or loss of packets at a single node and performance of the entire network. Furthermore, their proposal does not allow advance prediction of the cost for transmission of a single packet. It is generally believed that most consumers will not accept uncertainty in pricing.
Gupta, Stahl and Whinston have proposed prioritized service classes as a possible solution to congestion in the Internet in “Priority Pricing of Integrated Services Networks” in
Internet Economics
and elsewhere. However, their proposal introduces substantial overhead and undesirably results in low priority classes not getting any bandwidth when there is too much higher priority traffic.
In general, these proposals suffer from similar disadvantages in that they introduce an undesirable element of overhead complexity to the system and are overly dependent on per packet analysis. The primary disadvantages of focusing on per packet analysis is that most service degradation in the Internet is caused by unpredictable actions of others. Because packet switched network traffic is extremely bursty, it has proved difficult to accurately predict and optimize network traffic by predictive parameter adjustment.
A continuing roadblock to achieving an information superhighway on the Internet is the problem of traffic congestion and steadily deteriorating perceived quality of service as more and more users log on and ever increasing quantities of undifferentiated data are communicated across the Internet. As long as the contemporary network structure treats all data identically without an adequate mechanism for readily differentiating data transmissions on the basis of users' priority, the Internet will likely remain congested. Moreover, regulating network traffic to limit congestion should add minimal overhead and complexity to the network and it should be essentially predictable and widely applicable to a broad spectrum of communications.
SUMMARY OF THE INVENTION
In view of the foregoing, there is a need to regulate traffic over a packet switched network in a way that fairly prioritizes the traffic without unfairly penalizing low priority traffic or allowing excessive free riding. This invention provides a novel technique for regulating network traffic over packet switched networks such as the Internet. The present invention may be readily applied to any network data path, although it is presently contemplated that it may be most advantageously implemented in a backbone of a large data network such as the Internet.
In the present invention, the user is presented with the option of selecting a channel from a plurality of channels over which to send packets. These channels are only logically separate and part of the same physical network. The primary difference between the channels is the price charged to the user. Traffic management is provided largely by the users through their selection of a channel over which their packets are sent and received. The channels with higher prices would attract less traffic and would thereby provide a higher quality service.
The present invention is advantageously flexible in that either the sender or the receiver of information may be charged for transmission of a packet in a usage-sensitive priced packet switched network. Advantageously, charges may be incurred on the basis of packets sent or packets received.
Dividing the network into logical channels having graded costs will regulate traffic and limit congestion because users who perceive that the quality of service on a lower cost channel has degraded to an unacceptable level will, if they have the available resources, switch to a higher cost channel which, because of its higher cost, will have less traffic and hence less congestion. As each channel becomes unacceptably congested, the user will switch to progressively higher cost channels until the user achieves a subjectively acceptable balance of cost and perceived
AT&T Corp.
Kenyon & Kenyon
Kizou Hassan
Pezzlo John
LandOfFree
Technique for limiting network congestion does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Technique for limiting network congestion, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Technique for limiting network congestion will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2489650