An Approach to Flexible QoS Routing Active Networks Proceedings of the Fourth International Workshop on Active Middleware Services(AMS’02) 謝志峰 2002/11/14
Outline Introduction QoS Support in Active Networks AQR (Active QoS Routing) operation Simulation of AQR Conclusions
Introduction(1/2) Active Network(AN) are investigated since several years, attempting to satisfy the increasing needs of highly customizable protocol mechanisms. AQR:The paper combines the concept of AN with suitable QoS routing mechanisms to form a novel approach called “Active QoS Routing(AQR)”.
Three major concepts and terms in the context of AN will be referred to frequently in the remainder: Active Applications(AA): It denote user-provided communicating applications which make use of AN Execution Environment(EE):The runtime system available on an AN node is coined as EE NodeOS:The abstract machine on which all developer of customizations for an AN can rely is called NodeOS Introduction(2/2)
QoS Support in Active Networks Mechanisms which are usually associated with layers 3 or 4 : We find Active Congestion Control, which reduces the feedback delay for congestion control mechanisms by moving endpoint algorithms into the network. Mechanisms which transfer application layer functionality into the network : Intelligent dropping of packets that correspond to specific frames of a video stream.
1.The AQR sender calculates all non- cyclic paths to the destination form the cyclic paths to the destination form the link state routing table. link state routing table. 2. A probing packet carrying the QoS requirements, code for QoS calculation, the requirements, code for QoS calculation, the sender and receiver’s addresses and a list of sender and receiver’s addresses and a list of visited nodes is sent to each first hop of these visited nodes is sent to each first hop of these paths. paths. AQR (Active QoS Routing) operation (1/3)
3. Upon receiving an AQR probing packet, an AQR-compliant transit node executes the AA code,which Check if the minimum QoS requirements found in the packet can be met, Compares and updates the QoS data, Adds itself to the list of already visited nodes, and executes the code of the AQR sender, starting at step except that no probing packets are sent to the source or to any other already visited node. AQR operation (2/3)
4. Only packets which conform to the minimum QoS requirements reach the AQR receiver, QoS requirements reach the AQR receiver, where a list of valid paths is generated. After where a list of valid paths is generated. After a predefined period, the best path is chosen a predefined period, the best path is chosen and communicated to the sender and communicated to the sender AQR operation (3/3)
We performed two series of simulations with the “ns” network simulator. In all of our simulators, the nominal bandwidth of all links was 1.5Mbit/s, packet sizes of all packets including measurement packets were 500 bytes. Delay between probing packet ”waves” was set to approximately 2 RTTs, and we generally used a simulation time of 360 seconds. Simulation of AQR(1/10)
The goal of Figure 1 was to study the behaviour of delay based AQR in a somewhat realistic scenario. The sender was at node 9, the receiver was at node 45. Simulation of AQR (2/10)
One such result is depicted in fig.2. We chose this scenario because it shows a significant delay reduction(approx. 20%) despite a number of path changes. Simulation of AQR (3/10)
We chose to use a somewhat less realistic but more controllable scenario by mean of a 15- node topology, which is shown in fig.3. Using node 5 as a sender and node 13 as a receiver. We studied the behaviour of AQR both with (greedy) TCP background traffic and exponentially distrially UDP background traffic. Simulation of AQR (4/10)
Figure 4 shows the delay of a constant bit rate AQR stream with TCP background traffic. AQR based on bandwidth measurements alone not only increases the average delay but also jitter. Simulation of AQR (5/10)
In table 1(TCP background traffic) delay increased by approx. 9% in comparison with shortest path routing, jitter increased by 44%. In table 2(UDP background traffic) The throughput increased by 27% in comparison with shortest path routing.The average delay increased by 8% and jitter increased by 87%. Simulation of AQR (6/10)
We now focus on a mixture (called”AQR- new”) of both parameters, where a delay threshold limits the choice of paths. “AQR-old” denotes AQR solely relying on delay. Simulation of AQR (7/10)
There was no other drastic change in the delay or throughput results (see table 3) ; as could be expected, the average delay was notably 15% smaller than the average delay of shortest path routing (TCP background traffic). Simulation of AQR (8/10)
Unresponsive background traffic yields a different result, which is depicted in figure 7. (UDP background traffic) Simulation of AQR (10/10)
The main advantage of AQR-new with unresponsive background traffic lies in a throughput enhancement which was as high as 33% in our simulations. This enhancement is due to a smaller packet loss ratio. The average delay was reduced by 36%. Simulation of AQR
We have proposed AQR as an approach to combing Active Networks with QoS routing. In the variant finally proposed, AQR combines a consideration of both bandwidth and delay for finding optimal paths. This variant showed considerable improvements over shortest-path routing under various load combinations and characteristics. Conclusions
We can research related topic with Active Network. We can plan to consider multi-domain routing. We can research different topic with AQR. Future and related work