Performance Evaluation of QoS Routing Algorithms Karima Maalaoui, Abdelfettah Belghith, Jean-Marie Bonnin, Miled Tezeghdanti karima. [email protected], abdelfattah. [email protected]. tn, jean-marie. bonnin@irisafr, miled. tezeghdanti@enst-bretagnefr

Abstract New emergent distributed multimedia applications require end-to-end quality of service (QoS) guaranties and have stringent constraints on delay, jitter, cost, etc. Taking into account such needs in IP networks stands of utmost necessity. MPLS and traffic engineering provide indeed an adequate mean to establish constrained routes which satisfy application requirements especially in terms of bandwidth and delay. These routes are computed by running a QoS algorithm on the traffic engineering database maintained by an extended IGP routing protocol. In this paper, we first study and analyze three known QoS algorithms, namely SPF-TE, WSPF and DORA. This study is done through very detailed simulations using ONS, a simulator of the OSPF protocol, recently developed and integrated into the well known NS simulator. We show that while all these algorithms outperform clearly SPF, the algorithm usually used within OSPF, DORA algorithm stands slightly better than the two others. Then, we propose three different potential improvements of DORA and propose a new QoS algorithm termed WDORA that incorporates one these improvements. We show that WDORA outperforms DORA for different performance metrics. Key words: traffic engineering, constraints routing, QoS, OSPF, MPLS, WSPF, DORA, MIRA, NS.

1. Introduction The advent of new distributed multimedia applications requiring pseudo real time constraints such as video on demand, videoconference and telephony over IP led researchers, service providers and network operators to seriously consider quality of service policies. Several models were proposed to provide QoS in IP networks. The integrated services (IntServ) and the differentiated services (DiffServ) are among the architectures suggested for the QoS support within the IP network layer [9], [10]. These architectures are however insufficient for they lack means to control the traffic in IP networks. In practice, the traffic rather takes routes calculated by the underlying

0-7803-8735-X/05/$20.00©2005 IEEE

internal and external routing protocols which systematically provide the shortest paths. In the quest of a solution, traffic engineering has been recently attracting considerable attention in the research community. Traffic engineering performs the computation of constrained routes, the management of network resources and the load balancing. This is achieved by routers having detailed descriptions of the network state which can be ensured by OSPF-TE (OSPF with traffic engineering extension) [8]. These functionalities are very difficult to acquire in an IP network. Nevertheless, thanks to MPLS and the explicit routing, it became possible to take routes calculated by algorithms based on traffic engineering information and network load. The aim of this paper is to study route computation algorithms which offer the best results in the context of an IP/MPLS network implementing the OSPF traffic engineering extensions OSPF-TE. We first study three different known algorithms namely: SPF-TE (Shortest Path First with Traffic Engineering), WSPF (Widest Shortest Path First) [1] and DORA (Dynamic Online Routing Algorithm) [6]. The study is conducted through very detailed simulation using the ONS simulator (OSPF Network Simulator) [7]. ONS is an extension of the NS simulator consisting of detailed implementations of OSPF and OSPF traffic engineering extensions (OSPF-TE). We then present three different improvement of the DORA algorithm and propose a new QoS algorithm termed WDORA (Widest Dynamic Online Routing Algorithm). Traffic engineering is based on three modules: a module of traffic engineering information dissemination, a CSPF module (Constraint Based Shortest Path First) for path computation and a signaling module. Our work is integrated into the second module, namely the CSPF module. This model uses the OSPF-TE database which is responsible for gathering all the needed traffic engineering information. As a preliminary work, we have added the implementation of DORA and WSPF to the already existing CSPF module [2].

2. QoS routing algorithms The principal goal of a routing algorithm is to find a feasible path which uses efficiently the network resources. To optimize network performances, quality of service routing algorithms are virtually based on two different techniques. The first consists in selecting the path with minimal number of hops in order to reduce resource consumption. The second chooses the least loaded path to balance the network load. This optimization of the network utilization is not easily realizable by using a simple routing algorithm since these two goals can be opposite. For this reason, the development of a route computation algorithm must take into account more than just one aspect. A first simple method consists in using an algorithm that chooses only feasible paths with the least number of hops as a single routing criterion. This algorithm is called SPF-TE (Shortest Path First with Traffic Engineering) [8]. In [1] and [4], the authors proposed the WSPF routing algorithm (Widest Shortest Path First) based on the Bellman-Ford algorithm and which combines two criteria: hop number and available bandwidth. It selects the path with minimum hop number. If more than one path is selected then the one with maximum residual bandwidth MRB is chosen. Another routing proposal looks as the opposite of the WSPF where the first criterion is now taken to be the path with the maximum residual bandwidth and if more than one path is selected then the one with the smallest number of hops is chosen. This algorithm is called SWPF (Shortest Widest Path First) [4]. These algorithms have several disadvantages by choosing a path with a greater number of hops as in SWPF or a path with a critical allowance of bandwidth which could become a point of congestion as in WSPF. To avoid such a problem, other proposals use constraints which act to limit these disadvantages. We cite for example the DAP algorithm (Dynamic Alternative Path) [4] which is the WSPF along with a restriction on the hop number. Several proposals using MPLS networks to develop new path selection algorithms with QoS guarantees are proposed in the recent literature. Among these algorithms, we find MIRA (Minimum Inerference Routing Algorithm) [3], PBR (Profile Based Routing) [5] and DORA (Dynamic Online Routing Algorithm) [6]. MIRA takes into account the location of source and destination nodes. The main idea is that if routing is done obliviously with regards to the location of these sources and destinations of traffic then the chosen path may interfer with the routing of some future requests. The principal idea of MIRA is to forward traffic through the least interfering path with future requests. DORA is a dynamic online routing algorithm for the construction of bandwidth guaranteed paths in MPLSenabled networks. The main goal of DORA is to avoid

routing over links that have high potential to be part of any other path and have low residual available bandwidh. The operation of DORA lies into two stages. The first stage calculates the called path potential value (PPV) array associated with a source-destination pair. The second stage combines PPV with residual link bandwidth to form a weight value for each link of the path. This value is then used to compute a weightoptimized network path. The potential of a link being more likely to be included in a path than other links is characterized by an associated PPV. Formally, for each source-destination pair, we associate to each link an integer called the PPV with an initial value zero. Each source-destination pair (8, D) is associated with an array, PPV(S,Dj. When a path could be constructed over a link L for a given source-destination pair (81, D 1), we reduce PPV(Sl,Dlj(L) by 1. When a path could be constructed over the same link L for a different sourcedestination pair (82, D2), we increment PPV(Sl,Dlj(L) by 1. Since there are many paths between a given source-destination pair, we consider only disjoint paths. The computation of PPV arrays for each sourcedestination pair forms the first stage of the algorithm. In the second stage, we will remove all links with a residual bandwidth less than the required bandwidth. The PPV and the current residual bandwidth of each link are combined together to form the link weight. The content of the link weight is controlled by a parameter called BWP (BandWidth Proportion). Finally, we run Dijsktra algorithm to compute a weight-optimized path on the residual topology. 1.

WDORA We propose now a new algorithm termed WDORA (Widest Dynamic Online Routing Algorithm) which is based on the DORA algorithm and the following three improvements: e. In the computation of the set of disjoined paths between a source/destination pair (8, D), if there are more than one path with the same length having a common link, we choose rather the path with the largest bandwidth. For DORA, the choice is done arbitrarily. That is we propose to use WSPF algorithm instead of Dijkstra algorithm as in DORA. e In the computation of the PPV of a link L for a source/destination pair (S, D), we increment PPV(s,D)(L) by a value inversely proportional to the level of appearance of this link in the other paths between other source destination pairs. For example, if this link belongs to the shortest path between a source destination pair, we increase the PPV by 1 and if it belongs to the second shortest path we rather increase the PPV by 1/2 and so on. e In the computation phase, we propose to use WSPF algorithm instead of Dijkstra.

Currently, we have restricted our attention only to the third improvements. The two others will be included in a future work. Our proposal will then be compared to the aforementioned QoS algorithms.

3. Simulation 3.1. Comparison criteria To compare the various algorithms, the following performance criteria were used: Blocking probability: is defined as the ratio of the number of rejected connection requests to the total number of requests issued to the network. Response time: is defined as the sum of the path computation time and the time needed to actually establish this path if one does exist. Average length of the routes: is defined as the number of hops forming the selected route. This induces the effectiveness of the network resource utilization.

3.2. Parameters of the simulation

shortest path without taking into account traffic engineering parameters which may result in choosing paths not satisfying QoS requirements. We also notice that for the other algorithms, the blocking probability remains significant. This is explained by the choice of the interval time between two successive TE-LSA generations (MinLSInterval = 5 seconds) which means that traffic engineering information is outdated. These algorithms can consequently select routes which do not have actually required resources for the connection establishment. Whereas certain routes can satisfy the requests, they consider that not sufficient resources are available to establish the connections.

0.9

~0.8

:g.c

~ 0.6

C)

c

:s2

g

The network topology used in our simulation study is composed by 14 nodes and 28 links having a capacity of 4.8 MBps and 1.2 MBps. We have four distinguished pairs of source destination: (SO, DO), (S1, D1), (S2, D2), and (S3, D3). This is the same topology used in the study made to show performances of MIRA [3] compared to SPF and WSPF. This same topology was also used in [6] to show performances of DORA compared to MIRA and SPF. In our simulation, we studied the various performance criteria mentioned as a function of the load submitted to the network. We use the same definition as the one presented in [7] which is given by: Load = (Lb/C)*(A Ill) where: A = LsAs is the network cumulative connection request arrival rate with As representing the class s connection request. Request arrivals are assumed to be Poisson. L represents the average selected route length. b represents the average bandwidth for requested connections. Throughout the simulation, we fix b to 1.1MBps. C represents the total network capacity equal to the sum of all links capacities. IIIl represents the average connection duration.

4. Comparison of the algorithms 4.1. Blocking probability Figure 1 portrays the blocking probability in the computation and signalling phases versus the network load. We notice that all the algorithms based on traffic engineering extensions (SPF-TE, WSPF, DORA, WDORA) have a blocking probability much lower than that of the SPF algorithm. Indeed, SPF chooses the

0.7

0.5

iii 0.4 0.3

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Load (MinLSlnterval = 5 5)

Figure 1. Blocking probability

While the different QoS algorithms present very close behaviour, WDORA outperforms them slightly. This can be explained by the fact that WDORA selects routes having the largest residual capacities which results in routes having a greater probability of being still feasible for future requests. In addition, link costs in WDORA are different from those used in the case of WSPF. In fact, the new costs for WDORA take into account the degree of link memberships to other possible paths and the link bandwidths. WDORA maximizes then the percentage of acceptance of future demands by minimizing the interference.

4.2. Response time Figure 2 shows the response time as a function of the cumulative network load. At low workloads, SPFTE is the most powerful among all algorithms. For high workloads, it becomes less powerful than WSPF. Moreover, algorithms other than SPF-TE are ordered from more powerful to least powerful, as follows: WSPF, WDORA and DORA. This is explained by the average length of the routes as shown in figure 3. The route average lengths are ordered in perfect conformance for the response time. The later is indeed directly proportional to the route average length.

others. SPF always chooses the shortest path with no consideration to link capacities and traffic engineering information.

0.08 0.075 0.07

5. Conclusion and perspectives

~ 0.065

~

0.06

t/)

S 0.055 Q. t/)

~

0.05 0.045 0.04 0.035 L - - - - - - - L _ - - - - ' - _ - - - - I . . . - _ - - - l . . . . - _ . . . . L . . . . . - _ L - - - - - - - L _ - - - - ' - _ - - - - l 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Load (MinLSlnterval = 5 5)

Figure 2. Response time

4.3. Average length of the route Figure 3 represents the average route length for accepted connections as a function of the cumulative network load. Among the algorithms with traffic engineering, we notice that SPF-TE stands the best. In fact, SPF-TE uses only optimal routes which check the needs for traffic engineering. Moreover, the cost for SPFTE is the number of hops, which is not the case for DORA that has the longest routes even if it uses SPF for routes computation. This is explained by the fact that the link costs are computed according to bandwidth and PPV parameter representing the interference between paths. The same remark explains the fact that the curve of WDORA is above WSPF. As for the difference between DORA and WDORA and between SPF-TE and WSPF, we note that since WSPF and WDORA choose the broadest among the shortest routes. These routes will be then able to satisfy other future requests for connections. SPF-TE and DORA, however, can choose shorter routes whose bandwidths are consumed by just one or few connections. 3.9 spITE wspf dora05 wdora05 spf

3.8

J:

C,

3.7

c

3.6

!~

3.5

.!

e

CI> 3.4

C)

e 3.3 CI>

~

3.2 3.1

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Load (MinLSlnterval = 5 5)

Figure 3. Average route length

For traditional SPF algorithm, there is obviously a very great difference in route length compared to the

In this paper, we studied and compared performances of SPF, SPF-TE, WSPF and DORA routing algorithms. We showed that DORA does not well improve the routing performance. WSPF is an interesting algorithm that provides a clear improvement to the performances of SPF algorithm, yet remains simple to implement. We proposed a new routing algorithm WDORA which gives adequate results for all chosen performance criteria and outperforms DORA. This work concerned the intra-area traffic engineering. It will be interesting to extend it to multiareas traffic engineering. This study may also be augmented by considering the number of paths requiring a re-routing. When a link fails, all the affected paths must be re-routed, yet the re-routing itself can fail because of resources insufficiency. References [1] R. GUERIN and al. QoS Routing Mechanisms and OSPF Extensions. In Proceedings of the Globecom, 1997. [2] K. MAALAOUI. Evaluation des performances des algorithmes de routage a QoS. Master Thesis, ENSITunis/ENST-Bretagne, June 2004. [3] M. KODIALAM; T. V. LAKSHMAN. Minimum Interference Routing with Applications to MPLS Traffic Engineering, In Proceedings of the IEEE Infocom, 2000. [4] P. STEEKISTE; Q. MA. On Path Selection for Traffic with Bandwidth Guarantees, in Proceedings of the IEEE International Conference of Network Protocols, October 1997. [5] S. SURI; M. WALDVOGEL; P. R. WARKHEDE. Profile-based routing: A new framework for mpls traffic engineering, In Quality of Future Internet Services, Lecture Notes in Computer Science 2156, September 2001. [6] W. SZETO; R. BOUTABA; Y. IRAQI. Dynamic Online Routing Algorithm for MPLS Traffic Engineering, International Conference in Networking, 2002. [7] M. TEZEGHDANTI. Ingenierie de trafic intra-aire et multi-aires pour les reseaux dorsaux IP. Thesis, IRISA/INRIA - ENST Bretagne, December 2003. [8] G. Apostolopoulos, D. Williams, S. Kamat, R. Guerin, A. Orda and T. Przygienda, RFC 2676, QoS Routing Mechanisms and OSPF Extensions, August 1999. [9] R.Braden, D. Clark and S. Shenker, RFC 1633: Integrated Services in the Internet Architecture: an Overview, June 1994. Status: INFORMATIONAL. [10] S. Blake, D. Black, M. Carlson, E. Davies, Z. Wang and W. Weiss. RFC 2475 An Architecture for Differentiated Services, December 1998.

Performance evaluation of QoS routing algorithms - Computer ...

led researchers, service providers and network operators to seriously consider quality of service policies. Several models were proposed to provide QoS in IP ...

905KB Sizes 1 Downloads 382 Views

Recommend Documents

Performance evaluation of QoS routing algorithms
which satisfy application requirements especially in terms ofbandwidth and ... led researchers, service providers and network operators to seriously consider ...

Performance Evaluation of a QoS-Aware Framework for ...
good response times, by selecting different replicas to service different clients concurrently ... and responsiveness of the replicas by monitoring them at run- time.

Performance Evaluation of Ad Hoc Routing Protocols ...
ABSTRACT: An ad hoc network is a collection of wireless mobile nodes dynamically forming a temporary ... ireless networking is an emerging technology that.

Performance Enhancement of Routing Protocol in MANET
Ghaziabad, U.P., India ... Service (QoS) support for Mobile Ad hoc Networks (MANETs) is an exigent task due to dynamic topology and limited resource. To support QoS, the link state ... Mobile ad hoc network (MANET) is a collection of mobile devices,

TEACHER PROFESSIONAL PERFORMANCE EVALUATION
Apr 12, 2016 - Principals are required to complete teacher evaluations in keeping with ... Certification of Teachers Regulation 3/99 (Amended A.R. 206/2001).

Modeling and Performance Evaluation with Computer ...
Book synopsis. Queueing Networks and Markov Chains Critically acclaimed text for computer performance analysis--now in its second edition The Second ...

Performance Enhancement of the Temporally-Ordered Routing ...
dent on the services provided by the Internet MANET Encapsulation Protocol. (IMEP) to effectively carry out its three main ...... radio propagation properties of the mobile nodes are modelled after the Lucent. WaveLan direct sequence spread spectrum

CDOT Performance Plan Annual Performance Evaluation 2017 ...
48 minutes Feb.: 61 minutes March: 25 minutes April: 44 minutes May: 45 minutes June: 128 minutes 147 minutes 130 minutes. Page 4 of 5. CDOT Performance Plan Annual Performance Evaluation 2017- FINAL.pdf. CDOT Performance Plan Annual Performance Eval

PERFORMANCE EVALUATION OF CURLED TEXTLINE ... - CiteSeerX
2German Research Center for Artificial Intelligence (DFKI), Kaiserslautern, Germany ... Curled textline segmentation is an active research field in camera-based ...

Performance Evaluation of Equalization Techniques under ... - IJRIT
IJRIT International Journal of Research in Information Technology, Volume 2, Issue ... Introduction of wireless and 3G mobile technology has made it possible to ...

Performance Evaluation of Parallel Opportunistic Multihop ... - CiteSeerX
of the IEEE International Conference on Communications, Seattle,. WA, pp. 331-335 ... From August 2008 to April 2009, he was with Lumicomm Inc.,. Daejeon ...

Performance Evaluation of Equalization Techniques under ... - IJRIT
IJRIT International Journal of Research in Information Technology, Volume 2, Issue ... Introduction of wireless and 3G mobile technology has made it possible to ...

CDOT Performance Plan Annual Performance Evaluation 2017 ...
84% 159% 160% 30% 61% 81%. 113%. (YTD) 100% 100%. Whoops! There was a problem loading this page. Retrying... Whoops! There was a problem loading this page. Retrying... CDOT Performance Plan Annual Performance Evaluation 2017- FINAL.pdf. CDOT Performa

Performance Evaluation of Parallel Opportunistic ...
Department of Computer Science and Engineering, Dankook University, 152 ... Second, computer simulations are performed to verify the performance of the ...

Performance Evaluation of Curled Textlines ... - Semantic Scholar
[email protected]. Thomas M. Breuel. Technical University of. Kaiserslautern, Germany [email protected]. ABSTRACT. Curled textlines segmentation ...

Performance Evaluation of RANSAC Family
checking degeneracy. .... MLESAC takes into account the magnitude of error, while RANSAC has constant .... International Journal of Computer Vision, 6.

PERFORMANCE EVALUATION OF CURLED TEXTLINE ... - CiteSeerX
ABSTRACT. Camera-captured document images often contain curled .... CBDAR 2007 document image dewarping contest dataset [8] .... Ridges [5, 6] (binary).

Performance Comparison of Optimization Algorithms for Clustering ...
Performance Comparison of Optimization Algorithms for Clustering in Wireless Sensor Networks 2.pdf. Performance Comparison of Optimization Algorithms for ...Missing:

Experimental Performance Evaluation of a ...
packets among SW MAC, HW MAC, and Host-PC. The HW. MAC writes the packets received from the PHY into the shared-memory using Direct Memory Access ...

Performance Evaluation of Curled Textlines ... - Semantic Scholar
coding format, where red channel contains zone class in- formation, blue channel .... Patterns, volume 5702 of Lecture Notes in Computer. Science, pages ...

QoSBeeManet: a new QoS multipath routing protocol ...
Dec 6, 2010 - and real time applications in the MANET context,. • Need of a .... Good Packet Delivery Ratio (the best is DSR due to the cache routing policy).

PERFORMANCE EVALUATION AND ...
As uplinks of most access networks of Internet Service Provider (ISP) are ..... event, the TCP sender enters a slow start phase to set the congestion window to. 8 ...

QoS Routing for Wireless Ad Hoc Networks: Problems ...
Quality of service (QoS) provisioning is becoming a critical issue in designing wireless ad hoc net- works due to the necessity of providing multime- dia applications in such networks. These applications are typically delay-sensitive and have high ba

Multicast Routing and Its QoS Extension: Problems ...
Instead of sending a separate copy of the data to each individual ... for quality of service (QoS) fueled by emerging continuous ... cast/multicast data flows [2].