CA2234621A1 - Fair queue servicing using dynamic weights (dwfq) - Google Patents

Abstract

In a method of fair queue servicing at a queuing point in a multi-service class packet switched network, incoming packets are received in buffers and outgoing packets are scheduled by a weighted fair queue scheduler. Real-time information of buffer usage along with the minimum bandwidth requirement is used to dynamically modify the weights of the weighted fair queue scheduler.

Description

CA 0223462l l998-04-l4 Fair ~ueue Servicing using dynamic weights (DWFQ) This invention relates to the field of telecornmunications, and more particularly to a method of fair queue servicing in asynchronous data networks, such as Asynchronous Transfer Mode (ATM) networks or more generally any packet switched network that supports more than one class of service.
The use of ATM by a continually increasing number of applications is driving a requirement to increase the number of service classes and to allow more flexibility in the service offerings To support the application requirements, the ATM Forum is adding new service categories in new releases of ATM specifications. Furtherrnore, network providers are looking for the flexibility of defining multiple service classes for a given service categorv. The service classes are differentiated by their Quality-Of-Service requirements (QoS). The QoS requirements are configurable in accordance ~,vith a bi-dimensional matrix describing loss and delay. The delay jitter is another factor which needs to be bounded for some service classes.
Previously~ three service categories were supported on an ATM network element, namely constant bit rate (CBR), variable bit rate (VBR) and unspecified bit rate (UBR) The CBR service is the only service that guarantees a bound on delay. It is used for time sensitive data~ such as voice and video.
These v arious services can be supported by traditional exhaustive round-robin queuing among two priority queues. However, this simple technique cannot be usedwhen the number of queues increases beyond two, because of the high potential ofstarvation for lo~-er priority queues. Furthermore, the exhaustive round robin can only guarantee bounds on delay and delay variation for the highest priority queue. The support of multiple service class in an ATM switching product or multiplexer requires a minimum of one queue per class.
A queue scheduling algorithm. Weighted Fair Queuing (WFQ), has been recently proposed in the literature (see S. Golestani, ~ self-clocked Fair Queuing scheme for broadband applications INFOCOM 1994. ~une 1994) 06~as/ss as:s3 F~g CA 02234621 1998-04-14 ~08 .. ~ 2 ~
l~is s~ n~ scheme allows any nLunber queues (ser~ice classes) to be 1.
senriced, while providmg f~ir a~d work conser~ng access to b~ndwidth One of the key features of WFQ is thdt the CDV (Cell Delay Variation) ~s boul.d~l for any sen~ice class, as long as it is giv~n a ...1..,.."-.., weight~
~ Thisplo~os~ scheme can be imple.. ,L d in ATM products. Howe~er, it has not been ~ f-~rrnin~d how to set the servicin~ wei~hts ~-ffi-jfntly to take into account the dyn~ ic~llly ~hs~nein~ bandwidth ~.~ c.-~ of each service class (COnn~t'Or addition/removal, ABR flow control, Early packet Discard).
An object of the inverLtion is to p~ovide a Lan.e~voll; that ensures that the wei~hts are set d~ u~;ately to ~ e the desired Quality of Ser~rice and modified in real-time to ensure that the dynarnic Allnc~ n of bandwidth across the classes is opLi~
ALcor~ g to the present invention there is provided a me~od of fair queue servicing at a queu ng point in a multi~service class packet :,wiL~ihed network, where~n in~ min~ packets are received in buffers and outgoing packets are scheduled by aweightedfairqueues~ d~-lPrclL~dL.~;~e~inthatrePl-time;llf.ill,.Aii-nofbufferusage alongwiththeln.. ~;.~.. ~,baIldwidth~ tisusedtody.,~.nieAllymodifythe weighi:s of the weighted f~ur queue 5f~h~ r so as to cause buf~er queue si e to move toward a ~r~f ~ .i.,~l target value.
Preferablythe...illi.~ ll.bandw~dthr~ .Lis~ tP-l duri~gco~,Lon ~r1rniccir~n control Ihe medlod is pal1icularly suitable for use in ATM r~tworks.
The DWF Q (D~c Weighted Fair Queuing) can be impl.omPn~rrl at any queuing point which cl~lJiLLa~3 senr~c~g between n queues (n>2).
The invention also provides a fair queue servic~ng ~ ">~ ~1 in a multi-serv~ce class packet switched rletwork, comr~c;n~ a weighted f~ir queuing controller, and buffer means for rece~vLllg ;nco~ n~ packets in queues, ~ LP. ;7e~1 m that fi~her c~ .. ,l., ;~P~ I
means for mor~itoril~g buffer usage for each queue, means for det. ~ninin~ the bandwidth reqUir~ of each class of ser~r~ce, and a service weights nn~ns~e~r for dyn~min~lly modifying ~e weights of said weîghted fair queuing con~oller means in response to said bufferusage and bandw~dth~i~4.ne.~

A~k~l~ED StlEET

CA 0223462l l998-04-l4 Preferably, the means for monitoring buffer usage a queue growth monitor which perforrns real-time ~?s1im~tion of the queue growth in said buffer means.
The invention will now be described in more detail~ by way of example only. withreference to the accompanying drawings, in which:-Figure 1 is a diagram depicting the high level queuing scheme at an ATM switch;
Figure ~ shows the inforrnation provided by the Queue Growth Monitor, Figure 3 illustrates the data flow between the key components of the system andthe action of the Service Weight Manager (SWM); and Figure 4 describes the process performed by the SWM.
Referring now to Figure 1, ATM cells 2 arrive at buffer 1 and are placed in queues 11,12,...1n From there the cells are passed to a weighted fair queuing unit 3. The buffer I
is also connected to a queue growth monitor 4, which in turn is connected to congestion control unit 5, congestion analyzer 6, and connection admission controller 8, which in turn is connected to SVC & PVC (Switched Virtual Circuit and Permanent Virtual Circuit) connection handling unit 7, and service class manager 10. Queue growth monitor 4, connection admission controller 8 and service class manager 9 are connected to service weights manager 9, which is connected to weighted fair queuing scheduler 3 The key element of the Dynamic Weighted Fair Queuing (DWFQ) scheme is the service weight manager (SWM) 9, which dynamically modifies the service weights to be used by the WFQ Scheduler 3 It uses real-time information from the service classmanager 10, the connection admission controller 8, and the Queue growth monitor 4.
The service class manager 10 configures the service classes. A service class is configured with a given value of delay (CTD - Cell Transfer Delay) and loss (CLR - Cell Ratio Loss) requirements. These parameters represent the maximum nodal delay and loss allowed in order to meet the end-to-end QoS requirements of the connection. The service classes are mapped into a priority table as exemplified in Table 1. The priority table is used later by the service weight manager to allocate rem~ining bandwidth. The priority table is updated when a service class definition is modified. The service class manager also dictates which traff1c descriptors are used to compute the miniml-m bandwidth required by a connection of a given class.
CTD
CLR 100 ~s 500 lls None 10-~ 4 7 9 Table I ~ Example of a Queue Service Priority Mapping.
The connection admission controller (CAC) 8 computes the miniml-m bandwidth required for each service class. The minimum bandwidth is updated each time a connection of a given class is established or discormected, based on its traffic descriptor.
Table 2 shows a typical example of which traffic descriptors that can be used tocompute the minimum bandwidth for each basic service category relative to the queue service rate (SR) The CAC 8 communicates the minimum Weight table to the SWM
every time the value of the minimum weights have changed by a factor of ~,.
Queue; Category min Wi Ql CBR (~ PCR) / SR
Q2 RT-VBR (~ SCR) / SR
Q3 NRT- (~ SCR) / SR
VBR

Table 2 ~ Example of minimum weight table The CTD is further taken into account in the target queue size (TQS) table, which is the maximum queue size allowed to limit the CTD An example of TQSi computation is shown in Table 3, for typical service categories A zero TQS indicates that the queue can grow without limitation This table is computed by the CAC.

Queuei Category TQSi Q I CBR CTD / min Wi Q2 RT-VBR CTD / min Wi VBR

Table 2 ~ Example of a target queue size computation.
The Queue Growth Monitor (QGM) 4 performs real-time çstim~tion of the queue growth every Ts cell slots (sampling interval~. The information provided by the Queue Growth monitor 4 to the SWM 9 consists of ~Qi, the Queue Growth Rate of output queue i during an interval of duration Ts, Qi, the length of output queue i at the sampling time, and Ai, the arrival rate during the same interval of time.
The corresponding parameters: queue size Qi, queue growth ~Qi, number of arrivals Ai are collected or each queue by the QGM 4 for each Ts interval. From these parameters, auxiliary parameters such as average arrival rate ~i and service rate ~Li can be -derived by the SWM 9:
: average arrival rate, ?~j = Aj Ts llj: average service rate, llj = Sj - Ts~ where Sj = Aj--~Qj is the number of cells served during Ts.
Figure 2 shows the information provided by the Queue Growth Monitor 4. Using this inforrnation for the CAC 8, the service class manager 9 and the queue growth monitor 4, the SWM computes the service eight for each queue i (Wj) to be used during the next sampling interval.
As can be seen in Figure 3~ which shows the data flow between the key components of the system and the action of the SWM 9, the queue weights. Wi, areupdated usin~ information provided by the Queue Growth monitor 4.
If ~ j denotes the arrival rate of cells in queue i in the coming Ts interval, then ideally, the target service rate l~ j can be calculated as: (~ } Ts = Qi--TQSj . This means at the end of ne~t Ts interval, the queue size Qi will reach the target queue size TQSi. On the assurnption that ~j remains 1~nr.h:~n~ed from ~j, the service weight W, = ~ T5 can be approxim:~tecl as W~ Ts--~j ~ T5 + Qi--TQSj = Aj + Qj--TQSi .
However, the assurnption on the stable arrival rate may not hold, and also the actual number of serviced cells Si could be less than Wi; therefore a more conservative approach is ~Qj > 0, then Qj+~Qj, the predicted queue size at the end of the next Ts interval, is used to calculate the target service rate and weight. That is:
Ts = Q~ + ~Q,--TQSj and Wj--~j ~Ts +Qj +~Qj--TQSi = Aj +Qj + /~Qj--TQSi The detailed algorithm performed by the service weights manager 9 is shown in Figure 4. The queue size Qj at the end of each interval Ts~ the nurnber of arrivals Aj during the previous inten,~al Ts~ and the change in queue size ~Qi are input at step 20. Step 21 ~etermines whether the queue growth is positive: if yes, the sel~ice weight Wj is conservatively adjusted to bring the queue size to TQSi at step 22; if no, the service weight Wj is adjusted to bring the queue si~e to TQSi at step 23. The difference i~W is deterrnined in step 24.
Step ~5 determines whether the shared weights pool is empty: if yes, Wi is set to min_Wi in step 26; if no, step 27 determines whether Ws 2 ~Wi: if yes, step 28 sets Wj =
min_Wj + AWi and Ws = Ws ~ ~Wj; if no, step 29 sets Wj= min_Wi ~ Ws and Ws = 0.
Step 30 runs through all the Wj in the ordered list L and step 31 updates the ~eight table used by the Weighted Fair Queuing scheduler 3.
The described technique complies with ITU and ATM Forum standards and can be applied to any switching equipment which supports more than a single service class of servlce.

Claims (5)

Claims:

1. A method of fair queue servicing at a queuing point in a multi-service class packet switched network, wherein incoming packets are received in buffers and outgoing packets are scheduled by a weighted fair queue scheduler characterized in that real-timeinformation of buffer usage along with the minimum bandwidth requirement is used to dynamically modify the weights of the weighted fair queue scheduler so as to cause buffer queue size to move toward a predetermined target value.

2. A method as claimed in claim 1, the minimum bandwidth requirement is extracted during connection admission control.

3. A method as claimed in claim 1, characterized in that said weights are also modified in accordance with real-time service class information.

4. A method as claimed in claim 1, characterized in that buffer usage is monitored by a queue growth monitor, which performs real-time estimation of the queue growth every sampling interval T5.

5. A method as claimed in any of claims 1 to 4, wherein said packet switched network is an ATM network and said packets are ATM cells.