IP TRAFFIC ENGINEERING USING OMP TECHNIQUE
1KÁROLY FARKAS High Speed Networks Laboratory
Department of Telecommunications and Telematics Budapest University of Technology and Economics, H-1117, Pázmány P. s. 1/D., Budapest, Hungary
Phone: +36 1 463 3110, Fax: +36 1 463 3107 E-mail: farkask@ttt-atm.ttt.bme.hu
ABSTRACT
These days we live through significant changes with regard to using of Internet that is equipped with very constrained traffic management solutions. The demand on using eligible and efficient traffic management and engineering methods is increasing. This kind of efforts hallmarks the protocols/methods being under research/standardization phase, such as MPLS-Traffic Engineering, OSPF-OMP, MPLS-OMP, etc. This paper discusses the application of OSPF-OMP and MPLS-OMP (OMP: Optimized MultiPath) technique for Traffic Engineering purposes in IP networks and reviews OMP simulation results.
Keywords: Traffic Engineering, Optimized MultiPath (OMP), Open Shortest Path First (OSPF), Multi- Protocol Label Switching (MPLS), OPNET
1 This research was supported by Telia Research AB 1 INTRODUCTION
The explosive growth of the Internet over the last few years has made the IP protocol suite the most predominant networking technology. Furthermore, the convergence of voice and data communications over a single network infrastructure is expected to happen over IP-based networks. However, traditional IP networks offer little predictability of service due to the Best Effort behavior, which is often unacceptable for applications such as telephony, or real-time multimedia applications as videoconferencing. Today the efficient use of network resources and QoS (Quality of Service) guarantees in traditional IP networks are more-and-more important so we have to get accustomed to the terms such as TE, MPLS, QoS...
In this paper I first briefly review the idea of Traffic Engineering, OSPF routing protocol, MPLS technology and OPNET simulation tool. Then I discuss the general issues of OMP dynamic load-sharing algorithm and its use with OSPF and MPLS. Afterwards I present my simulation results in case of a basic network topology
using OSPF, OSPF-ECMP, OSPF-OMP, MPLS and MPLS-OMP techniques, respectively and in the last section we summarize the results and future plans.
1.1 TRAFFIC ENGINEERING
Traffic Engineering (TE) is concerned with performance optimization of operational networks [1]. In general, it encompasses the application of technology and scientific principles to the measurement, modeling, characterization, and control of Internet traffic, and the application of such knowledge and techniques to achieve specific performance objectives.
TE methods in IP networks (Figure 1) can be classified according to the type of the network where they are used: B-E – Best Effort or QoS – Quality of Service networks. The idea of QoS networks today consists of two main aspects, such as DS (Differentiated Services) [2] and IS (Integrated Services) [3] networks. Different techniques
can be used in different type of networks so QoS routing (e.g., QoSPF – Quality of Service Open Shortest Path First) can be used in IS while dynamic load-sharing (e.g., OMP technique) can be applied in DS and B-E networks for TE purposes. MPLS can operate on all these types of networks and it provides a convenient platform for TE.
TE in IP networks QoS
IS QoSPF DS
B-E
OMP
MPLS
Figure 1. Traffic Engineering in IP networks The aspects of Traffic Engineering that are of interest concerning MPLS are measurement and control. A major goal of Internet Traffic Engineering is to facilitate efficient and reliable network operations while simultaneously optimizing network resource utilization and traffic performance. Traffic and resource oriented performance can be increased by among other things minimization of the congestion in the network. Adopting load-balancing policies can reduce congestion resulting from inefficient resource allocation. The objective of such strategies is to minimize maximum congestion or alternatively to minimize maximum resource utilization, through efficient resource allocation.
1.2 OSPF
OSPF (Open Shortest Path First) is an internet routing protocol [4]. It is classified as an Interior Gateway Protocol (IGP). This means that it distributes routing information between routers belonging to a single Autonomous System (AS). The OSPF protocol is based on link-state and SPF (Shortest Path First) technology.
In a link-state routing protocol, each router maintains a database describing the AS's topology. Each participating router has an identical database. Every individual piece of this database is a particular router's local state (e.g., the router's usable interfaces and reachable neighbors). The routers distribute their local states throughout the AS by flooding. The possible routes to reach the destination nodes are computed from the database using the SPF algorithm that provides the shortest paths between the endpoints of the network.
1.3 MPLS
MPLS (Multi-Protocol Label Switching) [5] integrates a label-swapping framework with network layer routing.
The basic idea involves assigning short fixed length labels to packets at the ingress to an MPLS cloud. The forwarding function of a conventional router involves a capacity-demanding procedure that is executed per packet in each router. MPLS simplifies the forwarding function in the routers by introducing a connection-oriented mechanism inside the traditionally connectionless IP networks so label switched paths (LSP) are set up for each route or path through the network in advance using an IGP (e.g., OSPF). Throughout the interior of the MPLS domain, the labels attached to packets are used to make forwarding decisions (usually without recourse to the original packet headers).
1.4 OPNET
OPNET (OPtimized Network Engineering Tools) [6]
is a discrete event simulation tool, which provides a comprehensive development environment supporting the modeling and simulation of communication networks and which contains data collection and data analysis utilities.
OPNET allows large numbers of closely spaced events in a sizeable network to be represented accurately. It uses a modeling approach where networks are built of nodes interconnected by links. Each node’s behavior is characterized by the constituent components. The components are modeled as a state-transition diagram. I used OPNET as my simulation environment.
2 OMP TECHNIQUE
I proposed to study Traffic Engineering issues in IP networks. TE is a vast research area and I focused my efforts to study OMP (Optimized MultiPath) technique, which is a dynamic load-sharing algorithm, together with OSPF and MPLS in Best Effort networks.
2.1 OSPF-OMP
OSPF may form multiple equal-cost paths between source-destination node pairs. In the absence of any explicit support to take advantage of this, a path may be chosen arbitrarily. ECMP (Equal Cost Multi-Path) technique has been utilized to divide traffic somewhat evenly among the available paths. However, an unequal division of traffic among the available paths is generally preferable. Routers generally have no knowledge of traffic loading on distant links and therefore have no basis to optimize the allocation of traffic.
OSPF-OMP [7] utilizes the OSPF Opaque LSA (Link State Advertisement) option to distribute loading information inside the network, proposes a means to
adjust forwarding and provides an algorithm to make the adjustments gradually enough to insure stability yet provides reasonably fast adjustment when needed.
2.2 MPLS-OMP
In an MPLS network MPLS ingress routers may establish one or more paths to a given egress to the MPLS domain. Load can be balanced across a complex topology using MPLS. It requires that the ingress router is capable of computing a hash with a sufficiently fine level of granularity based on the IP source and destination addresses and selecting a forwarding entry based on the outcome of the hash.
MPLS-OMP [8] is an extension to MPLS. It does require that the IGP be capable of flooding loading information across the network. At the MPLS ingress an algorithm is applied to select alternate paths where needed and adjust forwarding. Forwarding is adjusted gradually enough to insure stability yet fast enough to track long term changes in loading.
3 SIMULATION RESULTS
My main goal concerning TE was to implement and examine load-balancing policies using OSPF-OMP and MPLS-OMP technique in OPNET simulation environment.
3.1 TEST NETWORK TOPOLOGY
My basic test network is depicted on Figure 2.
64 Kbps
10 Mbps 10 Mbps
Figure 2. Test Network Topology
The router nodes represented IP-based gateways in case of OSPF simulations and ATM-based gateways in case of MPLS simulations. The terminals and the server
were connected to the nearest router nodes by Ethernet connections using links with 10 Mbps capacity while the capacity of the links between the routers was 64 Kbps.
The link costs were the same everywhere in the network.
We can see that Link_a and Link_b constituted the bottleneck of this test network.
The test network modeled an almost ideal IP network in which the possibility of packet loss was zero using infinite queues at the router nodes. The same traffic scenario was used in all the simulations so the terminals made connections to the server node and offered almost constant bit rate traffic. The amount of the offered traffic by Terminal_2, Terminal_1 and Terminal_3 were 35 Kbps, 23 Kbps and 12 Kbps, respectively. The duration of every simulation was 3 hours. Terminal_1 and Terminal_2 started to transmit the traffic towards the server at the beginning of the simulations while Terminal_3 started the transmission 2 hours later (see Figure 3).
Terminal_2 Terminal_1 Terminal_3
Figure 3. Generated Traffic Rates by the Terminals 3.2 OMP SIMULATION RESULTS
I have run five different test scenarios on my test network. In these tests I used OSPF, OSPF with ECMP, OSPF with OMP, MPLS and MPLS with OMP techniques, respectively. In each case I monitored the throughput on the critical links of the network (Link_a and Link_b). Beyond that, I examined the differences of OSPF-OMP and MPLS-OMP and compared them in case of some other network topologies.
3.2.1 OSPF AND MPLS SIMULATIONS
In the first and fourth simulations (when I used “pure”
OSPF and MPLS) I got the results (see Figure 4, Figure 5) which were expectable if the routing is based on shortest path calculation.
Link_a
Link_b
Figure 4. Throughput in case of OSPF
Link_a
Link_b
Figure 5. Throughput in case of MPLS
If the shortest paths between different source and destination pairs contain the same critical link then link overloading can occur. (In my test network the shortest paths between Terminal_1 – server and Terminal_2 –
server contained the same low capacity link (Link_a) which got overloaded.)
The throughput on Link_a was very high and almost exceeded the capacity of the link while it remained low on Link_b. (In case of MPLS the aggregate throughput of the links was higher due to the relatively big amount of ATM header traffic.) These scenarios effected inefficient usage of network resources (i.e., the links’ capacities).
3.2.2 OSPF-ECMP SIMULATION
In case of OSPF-ECMP simulation the traffic generated by Terminal_2 was shared evenly among the equal cost paths between router_5 and router_2 so the network could avoid overloading on the critical links (see Figure 6) and provided better utilization than in the previous simulations. However, this sharing solution is static so the aggregate traffic can reach such amount that could cause overloading.
Link_b Link_a
Figure 6. Throughput in case of OSPF-ECMP 3.2.3 OSPF-OMP AND MPLS-OMP SIMULATIONS In OSPF-OMP simulation the OMP algorithm (see Figure 7) controlled the throughputs on Link_a and Link_b. When the system sensed high load on the critical links, OMP was activated and it directed load dynamically from the most heavily loaded link toward less loaded links to achieve a steady state. In this state the most heavily loaded links of the network were approximately equally utilized. The convergence time of the algorithm depends on the degree of the load in the network and the settings of the OMP’s parameters.
Link_b Link_a
Figure 7. Throughput of OSPF-OMP
OMP algorithm works similarly in MPLS networks.
Comparing the resulted graph (see Figure 8) with the graph of OSPF-OMP we can see that its structure resembles the other one due to the simple topology of the test network. The curves follow the changes of loading and converge to a state where the load on the most heavily loaded links is equally distributed. These scenarios gave the best results concerning the usage of the links’
capacities in case of our basic test network.
Link_b Link_a
Figure 8. Throughput of MPLS-OMP
However, the attainable utilization improvement of network resources using OMP significantly depends on
the network topology and the offered traffic profile. Due to the differences of the alternative route selection mechanism, MPLS-OMP can provide better result than OSPF-OMP – in that case if MPLS-OMP finds more alternative paths. (OSPF-OMP uses the relaxed Dijkstra algorithm in which any next hop, that is closer in terms of costs to the destination node than the current hop, can be considered a viable next hop for multipath routing. This restriction is needed to avoid the formation of routing loops. On the other hand, MPLS-OMP performs a SPF calculation to find alternative paths after temporarily removing the most loaded links from the OSPF link state database, so this could give some other longer paths beyond the OSPF-OMP paths.)
router_+
Link_c
Figure 9. Extended Test Network Topology To picture the dependence of the two OMP algorithms on the network topology I extended my basic test network (see Figure 9). An additional router node (router_+) was added to the core routers with two conjunctive links. The capacity of these links was 64 Kbps and the link costs were three times greater than the other link costs. I added the throughput graph on Link_c to the previously used statistics.
Link_b Link_a
Link_c
Figure 10. Throughput of OSPF-OMP
Link_b
Link_a
Link_c
Figure 11. Throughput of MPLS-OMP
In this topology OSPF-OMP didn’t consider the new path between router_5 and router_2 to be a viable route for the traffic coming from Terminal_2 (see Figure 10) while MPLS-OMP used it, so we could achieve a better result with MPLS-OMP (see Figure 11).
4 CONCLUSIONS AND FURTHER WORKS Today the importance of efficient traffic management solutions in IP networks is more-and-more increasing. If
the offered traffic toward a destination is too high, link overloading can be caused in the network using only
„pure” OSPF, MPLS. OMP technique solves this problem by flooding loading information across the network and balancing load between the alternative paths. In this paper I examined some basic features of OSPF-OMP and MPLS-OMP techniques by implementing them in OPNET simulation environment. The simulation results back up my expectations that OMP decreases the load of the most loaded links and it increases the utilization of the network resources but the improvement significantly depends on the network topology and the offered traffic profile.
Furthermore, I could achieve the best improvement with MPLS-OMP and either of the OMP techniques provided better results in every network scenario than the “pure”
OSPF, OSPF-ECMP or pure “MPLS” techniques.
On the other hand, it is important to investigate OMP technique from some other aspects. My future plans contain the investigation of the stability and convergence time of OMP technique in case of different parameter settings and network topologies. Moreover, I would like to examine the overhead traffic generated by OMP, the average and moving average of the aggregate network throughput, the usability of OMP in more realistic networks topologies and traffic patterns and I would like to implement and measure OMP in a real MPLS network.
REFERENCES
[1] Requirements for Traffic Engineering over MPLS, IETF RFC 2702, www.ietf.org/RFC/RFC2702 [2] An Architecture for Differentiated Services, IETF
RFC 2475, www.ietf.org/RFC/RFC2475
[3] The Use of RSVP with IETF Integrated Services, IETF RFC 2210, www.ietf.org/RFC/RFC2210 [4] OSPF Version 2, IETF RFC 2328,
www.ietf.org/RFC/RFC2328
[5] Multiprotocol Label Switching Architecture, Internet Draft, www.ietf.org/internet-drafts/draft- ietf-mpls-arch-06.txt
[6] OPNET Online Manual
[7] OSPF Optimized Multipath (OSPF-OMP), Internet Draft, www.ietf.org/internet-drafts/draft-ietf-ospf- omp-02.txt
[8] MPLS Optimized Multipath (MPLS-OMP), Internet Draft, www.ietf.org/internet-drafts/draft-ietf-mpls- omp-01.txt
