Pavlin Radoslavov

Topology-informed Internet replica placement

Recently, several studies have looked into the problem of replicated server placement on the Internet. Some of those studies have demonstrated that there exists a replica placement algorithm that can perform within a factor of 1.1-1.5 of the optimal solution. However, this particular greedy algorithm requires detailed knowledge about network topology, and knowledge about expected client locations on the topology. One of these previous studies has also looked at topology-informed replica placement. They consider placing replicas at highly connected nodes in the Autonomous System level graph. In this paper, we extend their work by investigating the performance of topology-informed placement on Internet router-level topology. In our evaluation, we consider approximated policy-based paths, and examine the sensitivity of our results to different client placements. We find that topology-informed replica placement methods can achieve average client latencies which are within a factor of 1.1-1.2 of the greedy algorithm, but only if the placement method is designed carefully.

The MASC/BGMP architecture for inter-domain multicast routing

Multicast routing enables efficient data distribution to multiple recipients. However, existing work has concentrated on extending single-domain techniques to wide-area networks, rather than providing mechanisms to realize inter-domain multicast on a global scale in the Internet.We describe an architecture for inter-domain multicast routing that consists of two complementary protocols. The Multicast Address-Set Claim (MASC) protocol forms the basis for a hierarchical address allocation architecture. It dynamically allocates to domains multicast address ranges from which groups initiated in the domain get their multicast addresses. The Border-Gateway Multicast Protocol (BGMP), run by the border routers of a domain, constructs inter-domain bidirectional shared trees, while allowing any existing multicast routing protocol to be used within individual domains. The resulting shared tree for a group is rooted at the domain whose address range covers the groups address; this domain is typically the group initiators domain. We demonstrate the feasibility and performance of these complementary protocols through simulation.This architecture, together with existing protocols operating within each domain, is intended as a framework in which to solve the problems facing the current multicast addressing and routing infrastructure.

A comparison of application-level and router-assisted hierarchical schemes for reliable multicast

One approach to achieving scalability in reliable multicast is to use a hierarchy. A hierarchy can be established at the application level, or by using router-assist. With router-assist we have more fine-grain control over the placement of error-recovery functionality, therefore, a hierarchy produced by assistance from the routers is expected to have better performance. In this paper, we test this hypothesis by comparing two schemes, one that uses an application-level hierarchy (ALH) and another that uses router-assisted hierarchy (RAH). Contrary to our expectations, we find that the qualitative performance of ALH is comparable to RAH. We do not model the overhead of creating the hierarchy nor the cost of adding router-assist to the network. Therefore, our conclusions inform rather than close the debate of which approach is better.

A framework for incremental deployment strategies for router-assisted services

Incremental deployment of a new network service or protocol is typically a hard problem, especially when it has to be deployed in the routers. First, an incrementally deployable protocol is needed. Second, a study of the performance impact of incremental deployment should be carried out to evaluate deployment strategies. Choosing the wrong strategy can be disastrous, as it may inhibit reaping the benefits of an otherwise robust service, and prevent widespread adoption. Unfortunately, to date there has been no systematic evaluation of incremental deployment for such services. Our research work is focused on the second aspect, namely the performance impact of incremental deployment of router-assisted services. We take the first step to define a framework for evaluating incrementally deployable services, which consists of three parts: (a) selection and classification of deployment strategies; (b) definition of performance metrics; and (c) systematic evaluation of deployment strategies. As a case study for our framework, we evaluate the performance of router-assisted reliable multicast protocols. Although our framework is still evolving, our results clearly demonstrate that the choice of a strategy has a substantial impact on performance, and thus affirms the need for systematic evaluation of incremental deployment. Our case study includes two router-assisted reliable multicast protocols, namely PGM and LMS. We make several interesting observations: (a) the performance of different deployment strategies varies widely; for example, with some strategies, both PGM and LMS approach full deployment performance with as little as 5% of the routers deployed, but with other strategies up to 80% deployment may be needed to approach the same level; (b) our sensitivity analysis reveals relatively small variation in the results in mos

Incremental deployment strategies for router-assisted reliable multicast

Incremental deployment of a new network service or protocol is typically a hard problem, especially when it has to be deployed at the routers. First, an incrementally deployable version of the protocol may be needed. Second, a systematic study of the performance impact of incremental deployment is needed to evaluate potential deployment strategies. Choosing the wrong strategy can be disastrous, as it may inhibit reaping the benefits of an otherwise robust service and prevent widespread adoption. We focus on two router-assisted reliable multicast protocols, namely PGM and LMS. Our evaluation consists of three parts: 1)selection and classification of deployment strategies; 2) definition of performance metrics; and 3) systematic evaluation of deployment strategies. Our study yields several interesting results: 1)the performance of different deployment strategies varies widely, for example, with some strategies, both PGM and LMS approach full deployment performance with as little as 5% of the routers deployed; other strategies require up to 80% deployment to approach the same level; 2) our sensitivity analysis reveals relatively small variation in the results in most cases; and 3) the impact associated with partial deployment is different for each of these protocols; PGM tends to impact the network, whereas LMS the endpoints. Our study clearly demonstrates that the choice of a strategy has a substantial impact on performance