Adaptive multicast topology inference

Duffield, Nick; Horowitz, J.; Presti, Francesco Lo

doi:10.1109/infcom.2001.916660

Cited by 55 publications

(70 citation statements)

References 11 publications

(22 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we focus on inferring the network topology. Most tomographic approaches rely on probes sent from a single source in a tree topology [7][8][9][10][11][12][13][14][15] and feed the number, order, or a monotonic property of received probes as input to statistical signal-processing techniques.…”

Section: Related Workmentioning

confidence: 99%

“…The resulting M -by-N topology is not exact, but bounds were provided on the locations of the points where the two 1-by-N trees merge with each other. This approach also requires a large number of probes for statistical significance, similar to many other methods [7][8][9][10][11]. Compared to [1], our work is different in that (i) we assume perfect knowledge of the quartets, thus we identify the topology accurately; (ii) we focus on the efficiency of active learning, i.e., selecting and merging the quartets, which has not been studied before.…”

Section: Related Workmentioning

confidence: 99%

“…One body of work developed techniques for inferring 1-by-N (i.e., single-source tree) topologies using endto-end measurements [7][8][9][10][11][12][13][14][15]. Follow-up work [1][2][3] showed that an M -by-N topology can be decomposed into and reconstructed from a number of two-source, two-receiver (2-by-2) subnetwork components or "quartets".…”

mentioning

confidence: 99%

“…1 More specifically, we start from the problem of 2-by-N topology inference, which is an important special case and can then be used as a building block for inferring an Mby-N . Consistently with [1], we assume that a (static) 1-by-N topology is known (e.g., using one of the methods in [4,[7][8][9][10][11][12][13][14][15]23]). Then we query the quartet component by sending end-to-end probes between the two sources and the two receivers, and we learn its topology using some of the methods in [1,2,16,17,[24][25][26][27][28] 2 .…”

mentioning

confidence: 99%

See 3 more Smart Citations

Active Learning of Multiple Source Multiple Destination Topologies

Sattari

Kurant

Anandkumar

et al. 2014

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

Abstract-We consider the problem of inferring the topology of a network with M sources and N receivers (hereafter referred to as an M -by-N network), by sending probes between the sources and receivers. Prior work has shown that this problem can be decomposed into two parts: first, infer smaller subnetwork components (i.e., 1-by-N 's or 2-by-2's) and then merge these components to identify the M -by-N topology. In this paper, we focus on the second part, which had previously received less attention in the literature. In particular, we assume that a 1-by-N topology is given and that all 2-by-2 components can be queried and learned using end-to-end probes. The problem is which 2-by-2's to query and how to merge them with the given 1-by-N , so as to exactly identify the 2-by-N topology, and optimize a number of performance metrics, including the number of queries (which directly translates into measurement bandwidth), time complexity, and memory usage. We provide a lower bound, ⌈ N 2 ⌉, on the number of 2-by-2's required by any active learning algorithm and propose two greedy algorithms. The first algorithm follows the framework of multiple hypothesis testing, in particular Generalized Binary Search (GBS), since our problem is one of active learning, from 2-by-2 queries. The second algorithm is called the Receiver Elimination Algorithm (REA) and follows a bottom-up approach: at every step, it selects two receivers, queries the corresponding 2-by-2, and merges it with the given 1-by-N ; it requires exactly N − 1 steps, which is much less than all N 2 possible 2-by-2's. Simulation results over synthetic and realistic topologies demonstrate that both algorithms correctly identify the 2-by-N topology and are nearoptimal, but REA is more efficient in practice.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Active Learning of Multiple Source Multiple Destination Topologies

Sattari

Kurant

Anandkumar

et al. 2014

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

show abstract

“…In the third approach, the network topology is inferred by monitoring packet arrival patterns at receiving hosts [7], [8]. The multicast-based topology detection technique [7] refers to the packet loss pattern.…”

Section: Related Workmentioning

confidence: 99%

Ethernet Topology Detection from a Single Host without Assistance of Network Nodes or Other Hosts

Hasegawa

Jibiki

2009

IEICE Trans. Commun.

View full text Add to dashboard Cite

SUMMARYTopology information has become more important for management of LANs due to the increasing number of hosts attached to a LAN. We describe three Ethernet topology discovery techniques that can be used even in LANs with Ethernet switches that have no management functionality. Our "Shared Switch Detection (SSD)" technique detects the Ethernet tree topology by testing whether two paths in the network share a switch. SSD uses only general MAC address learning. By borrowing MAC addresses from hosts, SSD can be run from a single host. The second technique determines whether two paths between two pairs of hosts contain a switch. The third reduces the number of shared switch detections. Simulation showed that these techniques can be used to detect the Ethernet topology with a reasonable search cost. Examination on a real-world testbed showed that they could detect an Ethernet topology consisting of six hosts and two switches within one second.

show abstract

An improved algorithm for multicast topology discovery from end‐to‐end measurements

Tian

Shen

2006

Int J Communication

View full text Add to dashboard Cite

We present a new multicast topology inference algorithm called binary loss tree classification with hop count (HBLT). HBLT improves the previous algorithm of binary loss tree classification (BLT) not only in time complexity but also in misclassification probability and inference accuracy. The time complexity of HBLT is Oðl 2 Þ instead of Oðl 3 Þ required by BLT in the worst case, and Oðl Á log lÞ instead of Oðl 3 Þ by BLT in the expected case, where l is the number of receivers in the multicast network. The misclassification probability of HBLT decreases more quickly than that of BLT as the number of probe packets increases. For correct classification, the inference accuracy of HBLT is always 1, i.e. the inferred tree is identical to the physical tree, whereas that of BLT is dependent on the shape of the physical tree and inversely proportional to the number of internal nodes with single child. We also show through simulation that HBLT requires fewer probe packets to infer the correct topology and hence has a lower misclassification probability and higher inference accuracy than BLT.

show abstract

Adaptive multicast topology inference

Cited by 55 publications

References 11 publications

Active Learning of Multiple Source Multiple Destination Topologies

Active Learning of Multiple Source Multiple Destination Topologies

Ethernet Topology Detection from a Single Host without Assistance of Network Nodes or Other Hosts

An improved algorithm for multicast topology discovery from end‐to‐end measurements

Contact Info

Product

Resources

About