2. An Effect Scheme for Fixed-Length Tunnel Allocation in Hierarchical
2.2 Background and Related Works
This section focuses on the related works that consider multi-granularity traffic.
More flexible and cost-efficient allocation of capacity is required to satisfy the growing demand for bandwidth. A considered method that has been studied intensively is to bundle a group of consecutive wavelength channels together and switch them as a single unit on a specific route to reduce the required resources of intermediate cross-connects along the route [1-6]. In [2], merits of hierarchical or multi-granularity OXC (MG-OXC) were summarized such as small-scale modularity, reduced cross-talk, and the reduced of complexity. [3] showed that the number of ports required when grouping of consecutive lightpaths are applied to the network (excluding grouping the traffic from different source nodes to different destination nodes) can be significantly reduced, compared to a traditional OXC solution. In [4], which employs a two-stage scheme of waveband and wavelength, an integer linear programming (ILP) formulation and a heuristic are given that aim to group lightpaths with the same destination only, while in [5] both the ILP and heuristic were given to handle the more general case.
While [3]-[5] discussed aim to dimension the network resources given the set of lightpath requests to be established, in [1] a novel switching architecture, MG-OXC was proposed to minimize the blocking probability for the dynamic requests given the limited network resources. In [1], the tunnel-like passage created by the bundled wavelength channels was defined as a waveband/fiber tunnel. If any residual capacity is left in the tunnel, it can be used to accommodate future lightpath requests. On the other hand, when no lightpath is traversing the tunnel, it can be torn down to release the resources including the link and the wavelength-switching ports as well as the multiplexer and de-multiplexer along the tunnel path for the future use. To illustrate this, the network in Fig. 4 is considered. Each link is assumed to have ten wavelengths,
λ1 to λ10, and can be divided into two wavebands with λ1 to λ5 being waveband 1 and λ6 to λ10 being waveband 2. Assume that a tunnel set up already exists from B to D on waveband 1 which is traversed by a lightpath from A to D on λ1. If the next lightpath request is from B to D, the established tunnel can be utilized to reach D. On the other hand, if the original lightpath is dropped, the tunnel can be torn down and the resources dedicated for this tunnel can be released for future use.
B C
Fig. 4 A network which has a tunnel with a lightpath traversing it
2.3 Basic Assumptions and Tunnel Allocation Characteristics
This section characterizes the tunnel allocation problem. Each node is assumed to be equipped with sufficient wavelength conversion capability in the wavelength-switching layer. Therefore, a lightpath in the wavelength-switching layer can be converted into any other wavelength if necessary. The tunnels are restricted to traverse only on their shortest paths from their ingress to egress node thus increasing the efficiency network resource consumption. The tunnels are restricted to traverse only on their shortest paths from their ingress to egress node thus increasing the efficiency network resource consumption.
A tunnel can be allocated between a node pair, if there is free capacity on each link along its route. Note that for the waveband tunnel, it has to use the same waveband on each link along the route. To bring up an allocated tunnel, wavelength-switching ports are further required at the two ends of the tunnel.
Fig. 5 illustrates an example of two possible tunnel allocations if the length of A
E D
tunnels is restricted to two. Fig. 5(a) is part of the physical network. Four fibers are used for tunnel allocation, includingAB, BD, DC and CA. Fig. 5(b) and (c) show the two possible ways of tunnel allocation. The total traffic trend should be considered when deciding which tunnel set is suitable. For example, if most traffic is between node A and node D, the tunnel set in Fig. 5(b) is more suitable. If most traffic is between node B and node C, the tunnels are allocated in Fig. 5(c).
(a) (b) (c)
B B B
D
A D A
A D
C C
C
Fig. 5 (a) Physical links to allocate tunnels (b) and (c) Two possible tunnel allocation
2.4 Heuristics for Tunnel Allocation
We first briefly introduce Capacity-Balanced Static Tunnel Allocation (CB-STA) proposed in [1]. Then we present our heuristic Constant Length Weighted Tunnel Allocation (CLWTA) that aims to improve CB-STA.
Capacity-Balanced Static Tunnel Allocation (CB-STA)
CB-STA aims to allocate tunnels off-line before start serving the lightpath requests.
The process comprises three stages: (a) tunnel ingress-egress (I-E) pair selection, (b) tunnel allocation and (c) makeup process. In (a), a series of I-E pairs are selected sequentially for the tunnel allocation stage in (b). To select I-E pairs, CB-STA estimates the amount of traffic traveling through each node by routing a historical traffic matrix in the network. Then the nodes with maximal traffic going out and maximal traffic coming in are selected repeatedly for tunnel allocation. In (b),
CB-STA tries to allocate a tunnel for each I-E pair selected in (a). After (a) and (b), the makeup process (c) is performed to further utilize the remaining resources to fill the fiber- and waveband-switching layer with as many tunnels as possible.
The tunnel allocated at stage (b) is required to follow a tunnel length constraint which is set to the minimum integer that is larger than the average physical hop distance between each node pair in the network. This is because when the tunnel length is too small, although the short tunnels are flexible and easily utilized by most of the lightpaths, the wavelength-switching ports are used up easily since the wavelength-switching ports are required at the ingress and egress nodes of each tunnel.
When the tunnel length is too large, although wavelength-switching ports can be greatly saved, the tunnels may not be suitable for the requests since most of the lightpath requests are shorter than the tunnels. We observe that the I-E pairs selected in stage (a) of CB-STA does not consider the tunnel length constraint, therefore most of the tunnels are allocated at the stage (c), leaving the performance of CB-STA some space to be improved.
Constant Length Weighted Tunnel Allocation (CLWTA)
CLWTA is proposed to overcome the problem in CB-STA. CLWTA allocate tunnels off-line and is based on an auxiliary used to rate the preference of tunnel allocation for each node pair. The process comprises four stages: (a) construction of auxiliary graph, (b) weight calculation for edges in the auxiliary graph, (c) weighted auxiliary graph based tunnel allocation, and (d) makeup process.
(a) construction of auxiliary graph Let G(V, Ep) be the original topology where V denotes the set of nodes and Ep represents the set of all physical links connecting the nodes. The auxiliary graph G’(V, E’) is constructed by adding auxiliary links El between the node pairs that have their shortest physical hop length
follow the length constraint (i.e., E’ = Ep + El). The auxiliary links represent the potential tunnels that could be allocated on the network.
(a)
D B
A F
C E
(b
ig. 6 gives an example of construction of auxiliary graph where Fig. 6(a) is the original topology with the average hop dist
Network Link S ach auxiliar
)
B D
A F
C E
Fig. 6 An example of auxiliary graph
F
ance equal to two and Fig. 6(b) is the corresponding auxiliary graph, in which dashed links represent the auxiliary links.
(b) weight calculation for edges in the auxiliary graph The Weighted tate (W-NLS) [1] is applied to determine the weight of e y link in the auxiliary graph. The weight of an auxiliary link is the predicted loads for the two nodes at the ends of that link. The larger the weight of an auxiliary link, the higher priority the node pair for that link gains to be allocated tunnels.
s d
Fig. 7 An example of deriving the W-NLS for each link in the network
Fig. 7 gives an example of how the weights are derived. There are three shortest paths from node s to d. The load from s to d is assumed to be equally distributed on the three paths. The weight of each link traversed by the shortest paths is thus increased by one third of the load from s to d. The weight of all the auxiliary links can be derived by applying the above procedure for all the node pairs in the network.
(c) weighted auxiliary graph based tunnel allocation This stage applies a greedy approach to allocate a set of tunnels according to the weight derived in the previous stage. The auxiliary link in G’ with the maximum weight is first selected, and an attempt is made to allocate a fiber tunnel for this auxiliary link. If the fiber tunnel can be allocated successfully, the weight of the corresponding auxiliary link is
decreased by
δ , where Wi,j is the weight of the auxiliary link connecting
node i and j, L the number of directional links in the original network topology, FT the number of fibers dedicated for tunnel allocation in each directional link and D the length constraint. Otherwise, we try to allocate a waveband tunnel for this auxiliary link. If a waveband tunnel can be successfully allocated, the weight of this auxiliary
link is decreased by
If both fiber and waveband tunnels fail to be allocated, the weight of this auxiliary
link is set to 0. The above procedure is repeated until all of the weights of the auxiliary links in G’ are equal to or less than 0.
(d) makeup process This process is used to further utilize the remaining resource after stage (c). The tunnels allocated in this stage do not have to follow the length constraint.
The whole algorithm of CLWTA is summarized as follows.
Constant Length Weighted Tunnel Allocation (CLWTA)
Step1. Form the auxiliary graph by adding all possible tunnels to the physical network.
Step2. Compute weight for each possible tunnel by routing the traffic matrix on the auxiliary graph
Step3. Stop if the weight for each auxiliary link is smaller or equal to 0.
Step4. Try to allocate fiber tunnel for the auxiliary link with maximum weight. If successful, decrease the weight of this auxiliary link by δF
and go to Step 3. Otherwise, go to Step 5.
Step5. Try to allocate waveband tunnel for this auxiliary link. Decrease the weight of this auxiliary link by δB. Go to Step 3.
In CLWTA and CB-STA, a tunnel can be allocated if free link capacity on the route between the ingress and egress of the tunnel is available. An allocated tunnel needs to be further brought up to be utilized by lightpaths. When a tunnel is brought up, wavelength-switching ports are needed so that wavelengths can be group or de-group at two ends of the tunnel. The number of wavelength-switching ports consumed at each end of the tunnel so that the tunnel can be brought up is equal to the capacity (in wavelength) of that tunnel.
We also propose another heuristic, Port Constraint- Constant Length Weighted Tunnel Allocation (PC-CLWTA) with slight modification on CLWTA. In PC-CLWTA,
after a tunnel is allocated, wavelength-switching ports at the ingress and egress nodes of the tunnel are dedicated to the tunnel. That is, a tunnel can not be allocated if any on the two ends of the tunnel has insufficient wavelength-switching ports.
PC-CLWTA improves the performance when the wavelength-switching ports is significantly fewer than the resources in the fiber-switching and waveband-switching layers. The performances of schemes described above are evaluated in the following section.
2.5 Numerical Results
The topology we use is a 16-node network show in Fig. 8. We assume that each directional link has five fibers. Each fiber contains forty wavelengths which are divided into four wavebands with wavelength 1 to 10 in the first waveband, 11 to 20 the second, …, and 31 to 40 the forth. The traffic is uniformly distributed on all node pairs and each request is for a lightpath. The inter-arrival time between two requests is determined by an poisson distribution function with rate ρ, and the request holds the resources it traverses for a time period determined by an exponential distribution function with rate 1.
Let (F1)F(F2)B(F3)L stand for the experiment with F1 fibers for fiber-switching, F2 fibers for waveband-switching, and F3 fibers for wavelength-switching for each directional link. The following three combinations of switching type are examined:
1F1B3L, 1F2B2L and 2F2B1L
1
Fig. 8 The 16-node network for this simulation
0
Fig. 9 Number of allocated tunnels except makeup process among CLWTA, CB-STA and the ideal number (1F2B2L)
Fig. 9 compares the number of allocated tunnels when CB-STA and CLWTA are used without performing their makeup process. The maximum number of allocated fiber tunnels and waveband tunnels are
D
|Ep| is the number of directional links on the topology. The number of allocated fiber/waveband tunnels without makeup process in CB-STA is considerably smaller
than the maximum number. The reason is that most of the I-E pairs selected in CB-STA do not follow the tunnel length constraint.
(a)
2F2B1L
0 0.05 0.1 0.15 0.2
450 500 550 600 650
Load
Blocking Prob.
CB-STA relaxed CB-STA WTA
(b)
1F2B2L
0 0.05 0.1 0.15
1000 1100 1200 1300 1400
Load
Blocking prob.
CB-STA relaxed CB-STA WTA
(c)
1F1B3L
0 0.05 0.1 0.15
1800 1900 2000 2100 2200
Load
Blocking Prob.
CB-STA relaxed CB-STA WTA
Fig. 10Comparison of blocking probability vs. requests for CLWTA and CB-STA on the 16-node topology
Fig. 10compares different blocking probability of the CLWTA and CB-STA under different load ρ. The relaxed CB-STA relaxes the length constraint D in CB-STA. More specifically, in relaxed CB-STA, tunnels with lengths between D-1 and D+1 are permitted to be allocated. Therefore, more useful tunnels can be allocated in relaxed CB-STA than in CB-STA. The results show that CLWTA has the lowest blocking probability in all switching type combinations. The reason is that CLWTA takes the length constraint into account when allocating tunnels while in CB-STA and relaxed CB-STA, length constraint is not carefully considered in their I-E pair selection stage.
(a)
2F2B1L
0 0.05 0.1 0.15
500 550 600 650 700
Load
Blocking prob.
WTA PC-WTA
(b)
1F2B2L
0 0.05 0.1 0.15
1300 1400 1500 1600 1700
Load
Blocking prob.
WTA PC-WTA
(c)
1F1B3L
0 0.05 0.1 0.15
1800 2000 2200 2400 2600
Load
Blocking prob.
WTA PC-WTA
Fig. 11 Comparison of blocking probability vs. requests for CLWTA and PC-CLWTA on the 16-node topology
PC-CLWTA outperforms CLWTA when each node in the MG-OXC network has only limited wavelength-switching ports (in Fig. 11(a)). That is because tunnels in PC-CLWTA are only allocated between nodes that have sufficient wavelength-switching ports. The link capacity and wavelength-switching ports are more efficiently utilized since most of them are consumed by the auxiliary links with higher weights. However, when there are sufficient wavelength-switching ports, performance of PC-CLWTA is the same as CLWTA (i.e., performance curves of the two algorithms in Fig.11 (b) and (c) overlaps).
Chapter 3
Tunnel-based Protection Schemes in Hierarchical Cross-connect WDM Networks
3.1 Introduction
In this Chapter, we investigate the protection schemes for single-link failure in hierarchical cross-connect WDM networks. Multi-granularity optical cross-connect (MG-OXC), has been proposed to support hierarchical WDM networks. It is attractive for its scalability and cost reason. However, the protection problem has not been intensively studied in MG-OXC networks, which make the mass MG-OXC deployment the risk of large data losses once a link failure occurs.
Our work in this chapter thus aims to provide efficient protection schemes for MG-OXC networks. To provide protection for lightpath requests, an intuitive solution is to allocate tunnels by CLWTA mentioned in Chapter 2 and then find two link-disjoint lightpaths from source to destination for each incoming request.
Although the intuitive heuristic provides a protection solution for the MG-OXC networks, it does not consider the protection when allocates tunnels. The lack of protection consideration while allocating tunnels complicates the finding of link-disjoint lightpaths since two different tunnels on the logical topology may actually traverse the same physical link. Thus, we have pay additional attention to the overlapping of tunnels when finding link-disjoint path pair. Therefore, we propose a protection scheme named Tunnel Based Segment Protection (TSP) that takes protection into consideration when allocating tunnels. In TSP, a protection tunnel is always allocated simultaneously with a working tunnel. The channels dedicated for protection can be shared more easily. In addition, performance of the network is
improved since working and protection tunnels use the same wavelength-switching ports in the network with MG-OXC in which port resources are rare. Simulation is conducted to compare two algorithms in MG-OXC networks. The results show that TSP works well in the hierarchical cross-connect WDM networks.
The rest of this Chapter is organized as follows. In Section 3.2, we describe the background and related works of protection. Concept of the protection schemes with MG-OXC is presented in Section 3.3. Performance evaluation between different protection schemes is given in Section 3.4.
3.2 Background and Segment Protection Schemes
This section introduces basic protection techniques in the WDM networks, including link protection and path protection, and the concept of shared protection.
Besides, segment-based protection which takes advantages of both link and path protection is also introduced. Two proposed algorithms, Short Leap Shared Protection (SLSP) [14] and Protection with Multiple Segments (PROMISE) [15] are also presented here as examples of segment-based protection.
We first examine two basic techniques, link and path protection, for survivability in the WDM networks. For path protection, if a fiber failure occurs on the working path, the end nodes of the failed link detects the fault and sends a notification signals to the source and destination of the path to activate a switchover. The source then immediately sends a wake-up packet to activate the configuration of the nodes along the protection path and after receives a confirm message from the destination, traffic are switched over from the working path to the protection path. On the other hand, link protection scheme reroutes all affected traffic over the prescheduled paths between the two ends of the failed link.
Based on whether the sharing of network resources is allowed, a protection
scheme can be further categorized as dedicated protection or shared protection. In dedicated protection, different protection paths do not share any link in the same wavelength plane. In shared protection, multiple protection paths may pass common links and share the same wavelength with each other. For example, in Fig. 12 (a) two working paths, W1 (A→B) and W2 (E→D) do not go through the same physical link.
The protection path P1 can share the same wavelength in link C-B with P2. However, if W1 and W2 go through the same physical link, shown in Fig. 12 (b), the sharing of P1 becomes impossible since the failure of link B-D would interrupt both these two working paths.
B
(a)
(b)
Fig. 12 (a) Two working paths can share the same wavelength in link C-B (b) Two working paths that have common link can not share the same protection resources
Shared protection provides one advantage over dedicated protection by offering higher network utilization. If all the lightpaths in the network need to be protected, the
A W1 D
W2 F
P1
E C
P2
B D
W1
W2 F A
C E
P1
best resource utilization is 50% in dedicated protection. But in shared protection, since the protection paths of link-disjoint working paths can share the resources with each other, network utilization can be higher than dedicated protection.
Compared to link protection, recovery in path protection may be slower since failure notification signal has to reach the source node of the lightpath before restoration is initialized. On the other hand, link protection may use more resources than path protection since in link protection a protection path has to be reserved for each link. To compromise both protection techniques, each working path can be
Compared to link protection, recovery in path protection may be slower since failure notification signal has to reach the source node of the lightpath before restoration is initialized. On the other hand, link protection may use more resources than path protection since in link protection a protection path has to be reserved for each link. To compromise both protection techniques, each working path can be