US 6,469,991 B1 · Method for overload control in a multiple access system for communication networks

Abstract

In the method for overload control in a wireless communications network employing On-Demand Multiple Access Fair Queuing, if the downlink/uplink buffer occupancy of the network has exceeded a high threshold, the base station determines if this is caused by a specific remote host or by a group of remote hosts. If caused by a specific remote host, the base station normally sends a flow control signal to the remote host to prevent it from sending more data, but may alternatively elect to disconnect other remotes if the remote experiencing bad performance is of a higher priority. The base station may additionally reduce the bandwidth shares allocated to any remote that have indicated tolerance for a variable allocated bandwidth. If the measured frame error rates for many remote hosts are increasing, then the base station may elect to disconnect those remote hosts that permit service interruption in order that more bandwidth may be allocated to the remaining users. If a majority of all associated remote hosts experience high uplink frame error rates, the base station may instead send a signal to a wireless hub which can coordinate the actions of other access points. Short packets queued up for so long at the base station that they exceed the time-to-live value allocated will be thrown away. The base station may also or alternatively elect to disconnect some users of a lower priority or redirect them to other nearby base stations that have a lower load. In a particular embodiment, an uplink Frame Error Rate (FER), an average uplink bit rate, a burstiness factor of uplink traffic, and a packet loss rate are measured at the base station for each remote host. Similarly, a downlink Frame Error Rate is measured at each remote host and then each FER is sent to the base station. If an overload condition exits, connections with a Frame Error Rate that has exceeded a threshold for a specified time and that have indicated that their connections can be interrupted are disconnected. Other combinations of the possible actions are suitable, with the exact combination being determined by the base station depending on the particular congestion conditions observed in the network.

Claims17

Dependent claims

measuring at said base station an uplink Frame Error Rate, an average uplink bit rate, a burstiness factor of uplink traffic, and a packet loss rate for each of said remote hosts;
measuring at each of said remote hosts a downlink Frame Error Rate, an average downlink bit rate, a burstiness factor of downlink traffic from said base station, and a packet loss rate;
sending said measured downlink Frame Error Rate from each of said remote hosts to said base station;
detecting at said base station from said Frame Error Rates when a queue overload occurs at least one of said remote hosts or said base station;
if said queue overload is detected at one of said remote hosts, sending flow-control messages between said at least one of said remote hosts and said base station in order to control data flow;
and
if said queue overload is detected at said base station, discarding any of said packets at said base station that have experienced a delay exceeding a time-to-live threshold.

4 more elements

A method for overload control in a wireless communications network employing the On-Demand Multiple Access Fair Queuing protocol, said network having a base station and a plurality of remote hosts between which uplink and downlink packets flow, each of said remote hosts and said base station having a respective uplink and downlink queue for said uplink and downlink packets, the method comprising the steps, in combination, of:

assigning a priority to each of said remote hosts;
measuring at said base station an uplink Frame Error Rate, an average uplink bit rate, a burstiness factor of uplink traffic, and a packet loss rate for each of said remote hosts;
measuring at each of said remote hosts a downlink Frame Error Rate, an average downlink bit rate, a burstiness factor of downlink traffic from said base station, and a packet loss rate;
sending said measured downlink Frame Error Rate from each of said remote hosts to said base station;
detecting at said base station from said Frame Error Rates when a queue overload occurs at least one of said remote hosts or said base station;
if said queue overload is in said base station uplink queue, sending flow-control messages between said at least one of said remote hosts and said base station in order to control data flow, including at least one of the steps of:
stopping packet flow from at least one of said remote hosts having a Frame Error Rate that exceeds a frame error rate threshold for a specified time and that has previously indicated that disconnection is allowed;
stopping packet flow from at least one of said remote hosts having the lowest priority;
redirecting packet flow from at least one of said remote hosts having the lowest priority to a nearby base station;
and
reducing bandwidth share allocated to at least one of said remote hosts that has previously indicated that a variable allocated bandwidth is allowed;
if said queue overload is in said base station downlink queue, performing at least one of the steps of:
discarding any of said packets at said base station that have experienced a delay exceeding a time-to-live threshold;
and
stopping packet flow to said base station from other parts of said network;
if said queue overload is in said uplink queue at at least one of said remote hosts, performing at least one of the steps of:
discarding any of said packets at said remote host that have experienced a delay exceeding a time-to-live threshold;
and
stopping packet flow to said remote host from other parts of said network;
if said queue overload is in said downlink queue at at least one of said remote hosts, performing at least one of the steps of:
stopping packet flow from said base station to said remote host;
and
disconnecting said remote host if it has previously indicated that disconnection is allowed;
and
if a queue overload is detected at a large number of said remote hosts, sending flow control messages to a wireless hub requesting coordination of nearby base stations.

22 more elements

A method for overload control in a wireless communications network employing the On-Demand Multiple Access Fair Queuing protocol, said network having a base station and a plurality of remote hosts between which uplink and downlink packets flow, said base station and said remote hosts having respective uplink and downlink buffers, the method comprising the steps, in combination, of:

detecting at said base station when a queue overload condition occurs because of at least one of said remote hosts or said base station, said queue overload condition occurring when the occupancy rate of one or more of said buffers exceeds a high threshold;
and
performing at least one overload control step, said at least one overload control step being selected from the following:
if said remote hosts are of varying priority, disconnecting at least one lower priority one of said remote hosts;
if said remote hosts are of varying priority, redirecting at least one lower priority one of said remote hosts to a nearby base station having a lower load;
if said queue overload condition is caused by only one of said remote hosts, stopping packet flow from said remote host to prevent said remote host from sending more of said packets;
if at least one of said remote hosts has indicated that a variable allocated bandwidth is allowed, reducing bandwidth share allocated to at least one of said remote hosts that has previously indicated that a variable allocated bandwidth is allowed;
if said queue overload condition is caused by a large number of said remote hosts, sending flow control messages to a wireless hub requesting coordination of nearby base stations;
if at least one of said remote hosts has previously indicated that service interruption is allowed if a queue overload condition is detected, stopping packet flow from at least one of said remote hosts allowing service interruption;
and
if said queue overload condition is caused by said base station, discarding any of said packets at said base station that have experienced a delay exceeding a time-to-live threshold.

8 more elements

assigning a priority to at least one of said remote hosts;
measuring at said base station an uplink Frame Error Rate, an average uplink bit rate, a burstiness factor of uplink traffic, and a packet loss rate for at least one of said remote hosts;
receiving a measured downlink Frame Error Rate from at least one of said remote hosts at said base station;
detecting at said base station from said Frame Error Rates when a queue overload occurs at least one of said remote hosts or said base station;
if said queue overload is in said base station uplink queue, sending flow-control messages between said base station and at least one of said remote hosts in order to control data flow, including at least one of the steps of:
stopping packet flow from at least one of said remote hosts having a Frame Error Rate that exceeds a frame error rate threshold for a specified time and that has previously indicated that disconnection is allowed;
stopping packet flow from at least one of said remote hosts having the lowest priority;
redirecting packet flow from at least one of said remote hosts having the lowest priority to a nearby base station;
and
reducing bandwidth share allocated to at least one of said remote hosts that has previously indicated that a variable allocated bandwidth is allowed;
if said queue overload is in said base station downlink queue, performing at least one of the steps of:
sending at least one flow-control message from said base station directing discarding of any of said packets at said base station that have experienced a delay exceeding a time-to-live threshold;
and
sending at least one flow-control message from said base station directing stopping of packet flow to said base station from other parts of said network;
if said queue overload is in said uplink queue at at least one of said remote hosts, performing at least one of the steps of:
discarding any of said packets at said remote host that have experienced a delay exceeding a time-to-live threshold;
and
stopping packet flow to said remote host from other parts of said network;
if said queue overload is in said downlink queue at at least one of said remote hosts, performing at least one of the steps of:
stopping packet flow from said base station to said remote host;
and
disconnecting said remote host if it has previously indicated that disconnection is allowed;
and
if a queue overload is detected at a large number of said remote hosts, sending flow control messages to a wireless hub requesting coordination of nearby base stations.

21 more elements

Dependencies are inferred from the claim text.

Method for overload control in a multiple access system for communication networks

Life of this patent

Method for overload control in a multiple access system for communication networks

Life of this patent

Abstract

Claims17