P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and maintain all or most group members. Some schemes attempt to reduce forwarding state at non branched routers [20] 18] 5] but they mainly target networks with a large number of sparse groups. Some other schemes try to achieve state reduction by forwarding state aggregation at routers [17], 19] Thaler and Handley analyze the aggregatability of forwarding state in [19] using an input output filter model. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidthmemory tradeoff with simulations in [17] However, these state aggregation schemes ....

....by forwarding state aggregation at routers [17] 19] Thaler and Handley analyze the aggregatability of forwarding state in [19] using an input output filter model. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidthmemory tradeoff with simulations in [17]. However, these state aggregation schemes attempt to aggregate routing state after the distribution trees have been established, and they tend to change the state format maintained at routers, which is generally not desired by many service providers [6] Furthermore, the state aggregatability of ....

[Article contains additional citation context not shown here]

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....non branching router dedicated approaches have limitations because they are designed only for net This work was supported in part by grant from NSF and Cisco Systems. works with large numbers of sparse groups. Some other schemes try to achieve state reduction by forwarding state aggregation [10, 12]. Thaler and Handley analyze the aggregatability of forwarding state in [12] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidth memory tradeoff with simulations in [10] Both these works attempt to ....

....state aggregation [10, 12] Thaler and Handley analyze the aggregatability of forwarding state in [12] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidth memory tradeoff with simulations in [10]. Both these works attempt to aggregate routing state after the distribution trees have been established. Implementation of this per router state aggregation is very complex. In addition, it is still an open question how much aggregation can be achieved or whether this approach is applicable in ....

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....and QoS multicast provisioning when the number of simultaneous on going multicast sessions is very large. Recently, much research effort has focused on the problem of multicast state scalability. Some schemes attempt to reduce forwarding state by tunneling[14] or by forwarding state aggregation[10, 13]. Thaler and Handley analyze the aggregatability of forwarding state in[13] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidthmemory tradeoff with simulation in [10] Both these works attempt to ....

....forwarding state aggregation[10, 13] Thaler and Handley analyze the aggregatability of forwarding state in[13] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidthmemory tradeoff with simulation in [10]. Both these works attempt to aggregate routing state after this has been allocated to groups. Second, some other architectures aim to completely eliminate multicast state at routers [6, 11] using network transparent multicast, which pushes the complexity to the end points. Though most research ....

[Article contains additional citation context not shown here]

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....state reduction. Some architectures aim to completely eliminate multicast state at routers [13, 24] using network transparent multicast, which pushes the complexity to the end points. Some other schemes attempt to reduce forwarding state by tunneling[28] or by forwarding state aggregation[23, 27]. Apparently, less entries are needed at a router if multiple forwarding state entries can be aggregated into one. Thaler and Handley analyze the aggregatability of forwarding state in[27] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate ....

....entries can be aggregated into one. Thaler and Handley analyze the aggregatability of forwarding state in[27] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidth memory tradeoff with simulation in [23]. Both these works attempt to aggregate routing state after this has been allocated to groups. It is still an open question how much aggregation can be achieved or whether this approach is applicable at all in real systems. Previously we proposed and evaluated a novel scheme called aggregated ....

[Article contains additional citation context not shown here]

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....problem with IP multicast when the number of simultaneous on going multicast sessions is very large. Recently, significant research effort has focused on the problem of multicast state scalability. Some schemes attempt to reduce forwarding state by tunneling[12] or by forwarding state aggregation[8, 11]. Thaler and Handley analyze the aggregatability of forwarding state in[11] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidth memory tradeoff with simulation in [8] Both these works attempt to ....

....forwarding state aggregation[8, 11] Thaler and Handley analyze the aggregatability of forwarding state in[11] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidth memory tradeoff with simulation in [8]. Both these works attempt to aggregate routing state after this has been allocated to groups. Second, some other architectures aim to completely eliminate multicast state at routers [4, 9] using network transparent multicast, which pushes the complexity to the end points. In this report, we ....

[Article contains additional citation context not shown here]

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....at routers using network transparent multicast, which pushes the complexity to the end points of the network. However, these high level multicast implementations are not very efficient. Some other schemes attempt to reduce forwarding states by tunnelling [11] or by forwarding state aggregation [8, 10]. Tian s tunnelling scheme [11] has limitations because it is designed only for sparse networks. Thaler and Handley analyze the aggregatability of forwarding states in [10] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding ....

....designed only for sparse networks. Thaler and Handley analyze the aggregatability of forwarding states in [10] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding states and study the bandwidth memory tradeoff with simulations in [8]. Both these works attempt to aggregate routing states after the distribution trees have been established. Implementation of this per router state aggregation is very complex. In addition, it is still an open question how much aggregation can be achieved or whether this approach is applicable in ....

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....state reduction. Some architectures aim to completely eliminate multicast state at routers [5, 9] using network transparent multicast, which pushes the complexity to the end points. Some other schemes attempt to reduce forwarding state by tunneling [11] or by forwarding state aggregation [8, 10]. Apparently, less entries are needed at a router if multiple forwarding state entries can be aggregated into one. Thaler and Handley analyze the aggregatability of forwarding state in [10] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to ....

....entries can be aggregated into one. Thaler and Handley analyze the aggregatability of forwarding state in [10] using an input output filter model of multicast forwarding. Radoslavov et al. propose algorithms to aggregate forwarding state and study the bandwidth memory tradeoff with simulations in [8]. Both these works attempt to aggregate routing state after the distribution trees have been established. We propose a novel scheme to reduce multicast state, which we call aggregated multicast. The difference with previous approaches is that we force multiple multicast groups to share one ....

[Article contains additional citation context not shown here]

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

....environment used for the simulations, which are discussed and evaluated in Section V. Section VI concludes this paper. II. AGGREGATING MULTICAST GROUPS As of now, prior work on aggregating multicast routing information is limited. One relevant scheme is the leaky aggregation scheme proposed by [5], which groups adjacent multicast addresses into a prefix. This operation is similar to the unicast prefix aggregation as done in CIDR. Their scheme requires routers to do perform an expensive second longest matching prefix lookup on the multicast group address, besides the standard lookup on the ....

....randomized IMPRESS over the baseline due to the reduced state that is required at the nodes. In terms of the data distribution requirements that were introduced in II, the router memory efficiency is minimal since we use a hashed lookup scheme as compared to an actual memory lookup scheme as in [5]. The mechanisms that we use also conserve address space where the bitmask results showed that reasonable performance was possible even when using 50 of the bits. B. Analysis of Randomized IMPRESS In this section, we refocus our attention the randomized IMPRESS scheme. We study the impact of a ....

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical Report 99-697, Department of Computer Science, University of Southern California, July 1999.

....and hierarchical approaches and is simpler. Very little work has been done in the area of multicast state aggregation. Work in[24] proposes an interface centric model for aggregation. This approach, however, benefits from having a large number of group members, which does not apply in our case. [25] studies strict, pseudo strict and leaky prefix aggregations for wide area multicast routing. We show, however, unlike most previous studies, that the bit wise aggregation usually achieves better gains than prefix aggregation. VIII. Conclusions We have presented a novel approach to IP ....

P. Radoslavov, D. Estrin, R. Govindan, "Exploiting the Bandwidth-Memory Tradeoff in Multicast State Aggregation", USC-CS-TR99-697, Computer Science Department, USC, July 1999.

....layer of overlay networks would be required on top of the existing MBone (Multicast Backbone) overlay. III. AGGREGATING MULTICAST GROUPS As of now, there exists very little work on aggregating multicast routing information. The only direct comparison is the leaky aggregation scheme proposed by [11], which groups adjacent multicast addresses into a prefix. This operation is similar to the unicast prefix aggregation as done in CIDR [6] Their scheme requires routers to do perform an expensive second longest matching prefix lookup on the multicast group address, besides the standard lookup on ....

....lack of variability and achieves equivalent performance when compared to the baseline IMPRESS. In terms of the data distribution requirements that were introduced in II, the router memory efficiency is minimal since we use a hashed lookup scheme as compared to an actual memory lookup scheme as in [11]. The mechanisms that we use also conserve address space where the bitmask results showed that reasonable performance was possible even when using 50 of the bits. The protocol also does not rely on the use consecutive address chunks. B. Randomized IMPRESS (Further Experiments) In this section, ....

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical Report 99-697, Department of Computer Science, University of Southern California, July 1999.

....replaces multiple forwarding entries with a single one, if the entries have adjacent group address prefixes and matching incoming and outgoing interfaces. It has been shown that hierarchical address allocation and membership locality increase the degree of state aggregation [19] Leaky aggregation [15] trades bandwidth for state, and allows data to leak downstream to links without receivers. Tunneling One common observation is that multicast state scalability is most critical at the core or backbone routers, because intuitively, most multicast trees pass through these highly connected ....

P. I. Radoslavov, R. Govindan, and D. Estrin. Exploiting the Bandwidth-Memory Tradeoff in Multicast State Aggregation. Technical report, University of Southern California/ISI, 1999. Submitted for publication.

....can support. Thaler and Handley state that . in the long run, the biggest issue facing multicast deployment is likely to be the scalability of multicast forwarding state as the number of multicast groups increases. 21] Several recent proposals reduce multicast state through aggregation [17, 21], application level clustering [14, 25, 26] tunneling [5, 6] and non branching state elimination [22, 20] We briefly describe these works next. 1.2 Multicast State Reduction State aggregation Forwarding state aggregation [21] replaces multiple forwarding entries with a single one, if the ....

....[21] replaces multiple forwarding entries with a single one, if the entries have adjacent group address prefixes and matching incoming and outgoing interfaces. It is shown that hierarchical address allocation and membership locality increase the degree of state aggregation. Leaky aggregation [17] trades bandwidth for state, and allows data to leak downstream without receivers. Tunneling One observation is that multicast state scalability is more critical at the core or backbone routers. Intuitively, most simultaneous multicast trees pass through these highly connected routers to ....

Radoslavov, P. I., Govindan, R., and Estrin, D. Exploiting the Bandwidth-Memory Tradeoff in Multicast State Aggregation. Tech. rep., University of Southern California/ISI, 1999. Submitted for publication.

....arises. Simple evaluation shows that this control loop serves as an effective adaptation process. IV. RELATED WORK Multicast state aggregation [19] reduces router forwarding state by combining entries with matching outgoing interface sets. To tradeoff for more state reduction, leaky aggregation [16] allows low bandwidth sessions to flow down links without receivers. This research addresses the problem of scaling multicast forwarding state across applications, which is not considered in our paper. However, it does not reduce the amount of control messages necessary to maintain such state. A ....

RADOSLAVOV, P. I., GOVINDAN, R., AND ESTRIN, D. Exploiting the Bandwidth-Memory Tradeoff in Multicast State Aggregation. Tech. rep., University of Southern California/ISI, 1999. Submitted for publication.

....can also be implicitly generated based on the current time. For example, the clustering algorithm can artificially give more preferences for sessions that are occurring now or will soon start, so more bandwidth is assigned to the current sessions. 5 Related Work Multicast state aggregation [27, 35] combines entries in a multicast forwarding table when the outgoing interface sets of the entries match. Variants of this approach include trading data leakage for more reduction in entries. This approach handles the problem of scaling multicast forwarding state across applications, which is not ....

....clustering across applications. The idea is to allocate multicast addresses to applications based on similarity in their durations, data rates, and subscriber locations. It is also worthwhile to study the tradeoffs between solving their problem in an end to end manner and at the router level [27, 35]. Acknowledgments We are grateful to Tzi cker Chiueh, Adam Costello, Christophe Diot, Mark Handley, Emmanuel Lety, Manuel Oliveria, Angela Schuett, and Helen Wang for their thoughtful discussions and comments on this paper. We also thank the anonymous reviewers for their insightful feedback. This ....

Radoslavov, P. I., Govindan, R., and Estrin, D. Exploiting the BandwidthMemory Tradeoff in Multicast State Aggregation. Tech. rep., University of Southern California/ISI, 1999. Submitted for publication.

....multicast delivery will solve this problem. Using one multicast group per event type eliminates the subscriber list at the sender. However, how to scale multicast routing to large numbers of global groups, given a relatively small IP multicast address space, is still the subject of research [14] [20]. Second, because receivers do not explicitly poll for an event, they have to rely on a failure detection mechanism to know whether sources have failed or whether there are simply no events. The most often used failure detection mechanism is a heartbeat message [11] In this situation, however, ....

RADOSLAVOV, P. I., ESTRIN, D., AND GOVINDAN, R. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Tech. Rep. 99-697, Dept. of Computer Science, USC, Feb. 1999.

No context found.

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

No context found.

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

No context found.

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

No context found.

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

No context found.

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

No context found.

P. I. Radoslavov, D. Estrin, and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report, USC Dept. of CS Technical Report 99-697 (Second Revision), July 1999.

No context found.

P. Radoslavov, D. Estrin, , and R. Govindan. Exploiting the bandwidth-memory tradeoff in multicast state aggregation. Technical report 99-697, University of Southern California, Dept. of CS, July 1999.

No context found.

P. Radoslavov, D. Estrin, , and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report 99-697, University of Southern California, Dept. of CS, July 1999.

No context found.

P. Radoslavov, D. Estrin, , and R. Govindan. Exploiting the bandwidthmemory tradeoff in multicast state aggregation. Technical report 99-697, University of Southern California, Dept. of CS, July 1999.

No context found.

P. Radoslavov, D. Estrin, R. Govindan, "Exploiting the BandwidthMemory Tradeoff in Multicast State Aggregation",USC-CSTR, July99.

No context found.

P. Radoslavov, D. Estrin, R. Govindan, "Exploiting the BandwidthMemory Tradeoff in Multicast State Aggregation", USC-CS-TR99697, Computer Science Department, USC, July 1999.

No context found.

P. Radoslavov, D. Estrin, R. Govindan, "Exploiting the BandwidthMemory Tradeoff in Multicast State Aggregation",USC-CSTR, July99.

No context found.

P. Radoslavov, D. Estrin, R. Govindan, "Exploiting the BandwidthMemory Tradeoff in Multicast State Aggregation", USC-CS-TR99697, Computer Science Department, USC, July 1999.

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