F. Bauer and A. Varma. ARIES: A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm. IEEE J. on Selected Areas in Comm., 15(3):382--397, 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....of the VPN. A possible method of nding a resilient topology is to favour core nodes with larger degree. Work on routing for multicast has considered the problem of dynamic group membership and the corresponding degradation in the quality of the routing tree. An example is the ARIES algorithm [Bauer97] in which a part of the multicast tree is rearranged when it has deteriorated beyond a certain threshold. The deterioration is measured by a degradation factor which is the ratio of the cost of the modi ed tree to the cost of the corresponding optimal Steiner tree. The rearrangement is limited ....

Fred Bauer and Anujan Varma. ARIES: A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm. IEEE Journal on Selected Areas in Communications, 15(3):382-397, April 1997. (p 73)

....node using the shortest path. For each delete request from a current member, GREEDY deleted only leaf nodes. If this deletion created a nonmember leaf it also deleted the new leaf. This process was continued until no nonmember leaves remains. Another heuristic had been proposed by Bauer and Verma [33], the heuristic was called ARIES. ARIES was like GREEDY where it did the minimum necessary modifications to the existing tree for each add and delete request. For each add request, ARIES joined the new member to the existing tree by its shortest path to the tree. For each delete request, ARIES ....

F. Bauer and A. Verma "ARIES: A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm" , IEEE Journal of Selected Areas in Communications, vol.15, no.3, pp 382-397, April 1997.

....that the cost of the tree is minimum, and (ii) multicast trees degenerate over time in presence of highly dynamic group. The research plan includes: Protocols for tree maintenance: As mentioned earlier, in a highly dynamic group multicast trees degenerate over time. Tree rearrangement [28, 29, 30] maintains a part of the multicast tree. Another way of tree maintenance involves reconstructing the multicast tree and moving the members from the old tree to the new one. This type of tree maintenance reduces the cost of the multicast tree by incurring huge amounts of service disruptions to the ....

F. Bauer and A. Verma, "ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm," IEEE JSAC, vol. 15, pp.382-397, Apr. 1997.

....multicast tree which are most e#ected by the members joining leaving the group. Both these techniques e#ect a portion of the multicast tree and hence are categorized as local tree maintenance. Both categories of local tree maintenance have been extensively studied by researchers over the years ([26, 27, 28]) The two categories are discussed below: Members Join Leave: When a node wishes to join leave a multicast group, it sends a graft prune message towards the core of the multicast group. The message traverses till it reaches an ontree node of the multicast tree. During join, the graft message ....

....[2] This incremental change approach su#ers 6 because the quality (e.g. tree cost) of the tree maintained may deteriorate over time. This technique of modifying the tree incrementally is identified as tree rearrangement. Some of the common tree rearrangement algorithms are GREEDY [26] ARIES [27] and the CRCDM [28] All these algorithms work by monitoring the accumulated damage to the multicast tree within local regions of the tree, as members join leave the multicast group. 2.3.2 Global Tree Maintenance Tree Migration If the multicast group is highly dynamic, the core of the ....

F. Bauer and A. Verma, "ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm," IEEE JSAC, vol. 15, pp.382-397, Apr. 1997.

....chooses the least cost configuration among four possible ways to connect the new member to the three identified nodes. If more than one cheapest alternative exists, GSDM chooses the one with greatest geographic spread. In addition, for each delete request, it deletes the nodes only if it is a leaf [25]. The GREEDY heuristic attempts to disturb the tree as little as possible. The new member is connected to the nearest tree node using the shortest path. A delete request leads to deletion of only leaf nodes and any non member nodes in the path. EBA heuristic enforces bounds on the distance ....

Fred Bauer, Anujan Varma, "ARIES: A Rearrangeable Inexpensive Edge-based on-line Steiner algorithm", In IEEE JSAC, Vol. 15, No. 3, April 1997, pp. 382-397.

.... role in tree maintenance and core failure recovery (see Figure 2) A number of heuristics have been proposed in [10] for core selection and are evaluated in [11] Several algorithms have been proposed for multicast tree construction [12] 13] and the references in [1] 3] and tree maintenance [14], 15] Despite its importance, to the best of our knowledge, the tree migration problem is not known in the literature. In his PhD thesis Carlberg [16] described a very simple core migration strategy where new core and the old core reside in the same tree. This strategy does not result in tree ....

F. Bauer and A. Varma, "ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm," IEEE JSAC, vol.15, no.3, pp.382-397, Apr. 1997.

....is used to suitably modify the tree. Our algorithm aims to satisfy the delay constraints of all current group members, at the same time minimizing the cost of the constructed tree. We compare the performance of our algorithm, by simulation, with that of an off line Steiner heuristic; with ARIES [2], a recently published algorithm for on line update of unconstrained trees; and with the algorithm proposed in [6] for on line update of delay constrained trees. The simulation results indicate that our algorithm provides excellent costcompetitiveness that is better than that provided by the ....

.... has simultaneously addressed the twin issues of dynamic group membership and delay constraint satisfaction is the algorithm proposed in [6] Existing approaches Unconstrained Case: The on line multicast problem was first presented by Waxman [14] and has since been addressed in [8] 9] and [2]. Waxman, in the first paper, divides on line heuristics into two categories: those that allow reconfiguration (or rearrangement) of the tree and those which do not. Imase and Waxman in [8] have investigated bounds for both kinds of heuristics. They have also shown that no such finite bound exists ....

[Article contains additional citation context not shown here]

F. Bauer and A. Varma, "ARIES : A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm," IEEE JSAC, vol. 15, no. 3, pp. 382-397, Apr. 1997.

....algorithms proposed in this paper come under this category. Non Steiner approaches (such as those proposed or employed on the Internet, including DVMRP, PIM, and core based trees) are suitable for datagram environments such as the Internet in which the routes taken by multicast packets may vary [3]. For such environments, there is no point in emphasizing cost minimization and more importance is given to minimization of algorithm overhead [3] 7] However, Steiner tree heuristic approaches apply to virtual circuit environments such as ATM networks [3] In such networks, the route (virtual ....

.... DVMRP, PIM, and core based trees) are suitable for datagram environments such as the Internet in which the routes taken by multicast packets may vary [3] For such environments, there is no point in emphasizing cost minimization and more importance is given to minimization of algorithm overhead [3] [7] However, Steiner tree heuristic approaches apply to virtual circuit environments such as ATM networks [3] In such networks, the route (virtual circuit) selected for a certain connection is used to forward all packets of that connection. Hence, it makes sense to model the cost of a ....

[Article contains additional citation context not shown here]

F. Bauer and A. Varma, "ARIES : A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm", IEEE JSAC, Vol. 15, No. 3, pp. 382-397, Apr. 1997.

....to suitably modify the tree. Our algorithm satisfies the delay constraints of all current group members, at the same time aiming to reduce the cost of the constructed tree. We compare the performance of our algorithm, by simulation, with that of an offline Steiner heuristic as well as with ARIES [4], a recently published algorithm for on line updation of unconstrained trees. From the simulation results, we conclude that our algorithm provides excellent cost competitiveness, reduces changes in the multicast tree after each update, and performs favorably even when compared with unconstrained ....

....the multicast tree after each addition or deletion. In the following section, we will review some existing approaches for the on line Steiner problem. 4.1. 1 Existing Approaches The on line multicast problem was first presented by Waxman [33] and has since been addressed in [16] 34] 19] and [4]. Waxman, in [33] divides on line heuristics into two categories : those that allow reconfiguration (or rearrangement) of the tree and those which do not. In the latter case, as new nodes are added to or removed from a group, rearrangement of existing routes is not allowed, i.e. addition of ....

[Article contains additional citation context not shown here]

F. Bauer and A. Varma, "ARIES : A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm," IEEE JSAC, vol. 15, no. 3, pp. 382-397, Apr. 1997.

....is used to suitably modify the tree. Our algorithm aims to satisfy the delay constraints of all current group members, at the same time minimizing the cost of the constructed tree. We compare the performance of our algorithm, by simulation, with that of an off line Steiner heuristic; with ARIES [2], a recently published algorithm for on line update of unconstrained trees; and with the algorithm proposed in [10] for on line update of delay constrained trees. The simulation results indicate that our algorithm provides excellent cost competitiveness that is better than that provided by the ....

....it s shortcomings. We also indicate how our algorithm overcomes these drawbacks to provide multicast trees with lower average costs. 2. 1 Existing Approaches Unconstrained Case The on line multicast problem was first presented by Waxman [26] and has since been addressed in [12] 28] 14] and [2]. Waxman, in the first paper, divides on line heuristics into two categories : those that allow reconfiguration (or rearrangement) of the tree and those which do not. In the latter case, as new nodes are added to or removed from a group, rearrangement of existing routes is not allowed, i.e. ....

[Article contains additional citation context not shown here]

F. Bauer and A. Varma, "ARIES : A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm," IEEE JSAC., vol. 15, no. 3, pp. 382-397, Apr. 1997.

....problems as follows [2, 8] First, each network node has to maintain routing information per every multicast tree 1 . Second, prune messages should be periodically exchanged between the network nodes in order to maintain an updated version of each multicast tree. The protocols presented in [2, 4, 8] address these problems. A tree spanning the members of the group is established in the network, and only the nodes sitting on the data path of the tree are required to maintain routing information associated with the tree. When a new node joins the multicast group, all the nodes on the shortest ....

....of the tree are required to maintain routing information associated with the tree. When a new node joins the multicast group, all the nodes on the shortest path between the new node and some node on the tree (the pre determined core in [2] the rendezvous point(s) in [8] or the closest node in [4]) join the tree and update their multicast routing tables accordingly. 1.1 Tunnel based multicast When multicast becomes pervasive more techniques must be employed in order to avoid scaling problems. If thousands of trees are created, deleted and updated every minute, the processing burden on ....

[Article contains additional citation context not shown here]

F. Bauer and A. Varma. ARIES: a rearrangeable inexpensive edge-based on-line steiner algorithm. In INFOCOM, March 1996.

....approach suffers from that the quality of the multicast tree maintained may deteriorate over time in terms of, for example, the total tree cost. Several researchers addressed the multicast tree rearrangement issue, among which the edge bounded algorithm (EBA) 27] Bauer and Varma s algorithm [28], Narvaez s algorithm [29] and Sriram s algorithm [30] may have received the most attention. The main idea is to define and monitor certain damage index to the multicast tree as members join leave, and trigger tree rearrangement when the index exceeds certain threshold. 4 Multicast Routing ....

F. Bauer and A. Varma. ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm. IEEE JSAC, pages 382---397, April 1997.

....number of nodes that will be join. However, actual traffic estimation is difficult, and setting is a complicated task for network management. The other conventional dynamic multicast routing algorithms take different approaches. For example, some allow re routing of multicast connections [11, 12, 17]. 4 Dynamic Multicast Routing Algorithm using Predetermined Path Search Our algorithm is based on the greedy algorithm and is intended to have the same performance. It uses query and reply messages to obtain membership information, as does the greedy algorithm. Thus, this approach does not ....

F. Bauer and A. Varma, "ARIES: a Rearrangeable Inexpensive Edge-based On-line Steiner Algorithm," IEEE INFOCOM'96, pp. 361--368, March 1996.

....traffic to existing members of the multicast group. Most solutions adopt a compromise where changes to the tree are kept as localised as possible, and periodically the entire tree is re routed to try and maintain a close to optimal topology. Some examples of current research in this area include [18] and [19] In [18] an algorithm is presented that maintains a good (but not optimal) multicast tree (in terms of minimising the sum of the edge weights) without undue computational cost using Kruskal s shortest path algorithm. Rearrangements are triggered after a certain amount of deterioration ....

....members of the multicast group. Most solutions adopt a compromise where changes to the tree are kept as localised as possible, and periodically the entire tree is re routed to try and maintain a close to optimal topology. Some examples of current research in this area include [18] and [19] In [18], an algorithm is presented that maintains a good (but not optimal) multicast tree (in terms of minimising the sum of the edge weights) without undue computational cost using Kruskal s shortest path algorithm. Rearrangements are triggered after a certain amount of deterioration in the quality of ....

[Article contains additional citation context not shown here]

Fred Bauer and Anujan Varma. ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm. IEEE Journal on Selected Areas in Communication, 15(3):382--397, April 1998.

....21) have to be modified. On the other hand, the quality of a multicast tree maintained may degrade over time as group members dynamically join leave the tree, and a new multicast tree may have to be re constructed if the existing tree cannot make efficient use of network resources. Bauer et al. [1], Sriram et al. 13] and Narvaez et al. 9] proposed several heuristics for locally rearranging part of an existing multicast tree so as to strike a balance between the multicast service disruption and the quality of a multicast tree. The suitability of using these approaches in the proposed ....

F. Bauer and A. Varma. ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm. IEEE JSAC, pages 382--- 397, April 1997.

....the on line multicast heuristics into two types: the ones that do not allow rearrangement (nonrearrange able) of the multicast tree and those that allow rearrangement (rearrangeable) when the cost exceeds some limit. Several heuristics which approximate the OMCMT problem have been suggested [18, 4, 6, 10, 11, 19, 9], but none of these address supporting the delay requirements of multimedia applications. Following is a review of some of these heuristics. a) On Line Greedy Heuristic (OGH) This heuristic works as follows. In response to a join request, a node is added to the multicast tree using the shortest ....

....been shown that GSDM heuristic usually performs slightly better than OGH [11] GS(T ) X v2V z2T SP (v; z) # Gamma1 (1) where SP(v,z) is the shortest path between v and z. e) ARIES ARIES (A Rearrangeable Inexpensive Edge Based On Line Steiner Algorithm) is a rearrangeable heuristic [4]. ARIES performs a rearrangement of a region of the multicast tree when the number of modifications (join, leave) within that region reaches a threshold. A region is defined as the part of the multicast tree whose interior nodes are non stable nodes. A stable node is a node that has never been ....

F. Bauer and A. Varma. Aries: A rearrangeable inexpensive edge-based on-line steiner algorithm. IEEE Journal on Selected Areas in Communications, 15(3):382--397, April 1997.

.... In his work, Waxman partitioned the on line multicast heuristics into two types: those that do not allow rearrangement (non rearrangeable) of the multicast tree and those that allow limited rearrangement (rearrangeable) Several heuristics which approximate the OMCMT problem have been suggested [19, 4, 6, 11, 12, 20, 10]. Most of these heuristics either bound the delay or minimize the cost of the multicast trees but not both. The following is a review of some of these heuristics. 2.3.1 Least Delay Heuristic Least Delay Heuristic (LDH) is the only heuristic among the ones reviewed in this section that bounds the ....

....shortest path between v and z. 2.3. 6 ARIES ARIES (A Rearrangeable Inexpensive Edge Based OnLine Steiner Algorithm) is a rearrangeable heuristic that performs a rearrangement of a region of the multicast tree when the number of modifications (join, leave) within that region reaches a threshold [4]. A region is defined as the part of the multicast tree whose interior nodes are non stable nodes. A stable node is a node that has never been modified since the start of the tree or the last rearrangement of that region. The performance and the time complexity of ARIES algorithm depend on the ....

F. Bauer and A. Varma. Aries: A rearrangeable inexpensive edge-based on-line steiner algorithm. IEEE Journal on Selected Areas in Communications, 15(3):382--397, April 1997.

....algorithms for the CM problem; however, CM generalizes some well studied problems. In case S = D = V the CM problem reduces to the gossiping problem which arises in a large class of scientific computation problems [10] In case jSj = 1 and D V the CM problem reduces to the multicasting problem [1, 4, 5, 27] and in case D = V to the broadcasting problem, both problems have been well investigated because of their relevance in the context of parallel distributed systems [26] In particular the broadcasting and gossiping problems have been investigated under a varieties of communication models and ....

....as given in Figure 2, but with a different choice of the Steiner trees T S and TD in order to have the possibility of dynamically and efficiently modify them according to the sequence of requests. Several papers have recently considered the problem of efficiently maintaining dynamic Steiner trees [1, 5, 19]. It is easy to see that that any algorithm for the dynamic Steiner tree problem can be applied to obtain equivalent results for our CM problem. 6 Concurrent multicast without block concatenation In this section we consider the concurrent multicast problem under the hypothesis that each message ....

F. Bauer, A. Varma, "Aries: a Rearrangeable Inexpensive Edge--Based On--Line Steiner Algorithm, Proceedings of INFOCOM'96, 361--368.

....On a completely different track, the Steiner tree algorithm seeks to build an optimal tree involving the receivers of the multicast group. While the Steiner tree problem is known to be NP complete, a number of algorithms seek to approximate its performance with acceptable computational overhead [10, 12, 13, 9, 11]. For example, 10] and [12] use the shortest path Steiner tree heuristic to build a delay constrained minimum cost multicast trees and degreeconstrained minimum cost multicast trees respectively. However, these algorithms assume that the membership is stable. Unfortunately, as pointed out in ....

....11] For example, 10] and [12] use the shortest path Steiner tree heuristic to build a delay constrained minimum cost multicast trees and degreeconstrained minimum cost multicast trees respectively. However, these algorithms assume that the membership is stable. Unfortunately, as pointed out in [13], the dynamics of group membership causes a violation in the performance bound achieved by these algorithms. In order to address this problem, Aries [13] rebuilds the multicast tree (using the Steiner tree heuristic) after a threshold number of changes in the group. However, rebuilding the tree ....

[Article contains additional citation context not shown here]

F. Bauer, A. Varma, "ARIES: A Rearrangeable Inexpensive Edge-based On-line Steiner Algorithm" IEEE, 1996.

....Rebuilding a new tree in this way, however, may cause large disruptions for the current multicast sessions. This is especially unacceptable for real time multimedia applications. Most existing dynamic multicast routing heuristics consider only one link metric (link cost or link delay) 6] 7] [8] [9] 10] The only known existing dynamic multicast routing algorithm that considers more than one link metric is the WAVE algorithm [11] WAVE works as follows: when a new member v n joins an existing tree, vn contacts the source node s with a request Req. Starting from s, this request is ....

Bauer F. and Varma, A., ARIES: A Rearrangeable Inexpensive Edge-based On-line Steiner Algorithm, in proceedings of GLOBECOM'95, 1995

No context found.

F. Bauer and A. Varma. ARIES: A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm. IEEE J. on Selected Areas in Comm., 15(3):382--397, 1997.

No context found.

F. Bauer and A. Varma. "ARIES: A Rearrangeable Inexpensive Edge-based On-line Steiner Algor ithm," in Proc. IEEE INFOCOM, San Francisco, Mar. 1996, pp. 361--368.

No context found.

Fred Bauer and Anujan Varma, "ARIES: A Rearrangeable Inexpensive Edge-Based On-Line Steiner Algorithm," IEEE Journal of Selected Areas in Communications, vol. 15, no. 3, pp. 382--397, 1997.

No context found.

Bauer, F., Varma, A.: Aries: A rearrangeable inexpensive edge-based on-line steiner algorithm. IEEE Journal of Selected Areas in Communications (1995)

No context found.

F. Bauer, A. Varma, ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm, IEEE Journal of Selected Areas in Communications 15 (3) (1997) 382--397.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC