An efficient algorithm for load partitioning in multicore shared-Tree multicast

Koushik Sinha, Bidyut Gupta, Ansuman Bhattacharya

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Core Based Tree (CBT) multicast involves using a single shared minimum cost tree rooted at a designated core node, to handle multiple multicasts. As the core node represents a single point of failure, recent works have explored the use of multiple candidate cores to increase the fault-Tolerance of CBT multicasting. In [20], the authors proposed utilizing these candidate cores to not only improve fault tolerance, but also for load sharing of multicast traffic between the cores. The load partitioning approach in [20] utilizes the concept of pesudo-diameter for calculating the load to be offered by a multicast source to each of the candidate cores. In this paper, we present a new multicore based load partitioning algorithm, in which each multicast source node generates a fibonaaci sequence based partitioning of its data traffic load between the candidate cores. Through simulation, we have compared the performances of our proposed Fibonacci multicore load partitioned multicast (FMLPM) algorithm with the load partitioning scheme proposed in [20]. For the mean makespan of multicast packet delay, our simulation results show that our proposed FMLPM algorithm provides on an average a 3.5% improvement over the load partitioning algorithm proposed in [20]. The FMLPM algorithm has an added benefit of not requiring any floating-point operation for calculating the load to be offered to each candidate core by a multicast source node.

Original languageEnglish
Title of host publicationProceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017
PublisherThe International Society for Computers and Their Applications (ISCA)
Pages149-153
Number of pages5
ISBN (Electronic)9781943436064
Publication statusPublished - 2017
Event32nd International Conference on Computers and Their Applications, CATA 2017 - Honolulu, United States
Duration: 20 Mar 201722 Mar 2017

Other

Other32nd International Conference on Computers and Their Applications, CATA 2017
CountryUnited States
CityHonolulu
Period20/3/1722/3/17

Fingerprint

Fault tolerance
Multicasting
Costs

Keywords

  • Core based tree
  • Fibonacci
  • Golden ratio
  • Load partitioning
  • Multicore multicast
  • Packet delay
  • Pseudo diameter

ASJC Scopus subject areas

  • Computer Science Applications

Cite this

Sinha, K., Gupta, B., & Bhattacharya, A. (2017). An efficient algorithm for load partitioning in multicore shared-Tree multicast. In Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017 (pp. 149-153). The International Society for Computers and Their Applications (ISCA).

An efficient algorithm for load partitioning in multicore shared-Tree multicast. / Sinha, Koushik; Gupta, Bidyut; Bhattacharya, Ansuman.

Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017. The International Society for Computers and Their Applications (ISCA), 2017. p. 149-153.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sinha, K, Gupta, B & Bhattacharya, A 2017, An efficient algorithm for load partitioning in multicore shared-Tree multicast. in Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017. The International Society for Computers and Their Applications (ISCA), pp. 149-153, 32nd International Conference on Computers and Their Applications, CATA 2017, Honolulu, United States, 20/3/17.
Sinha K, Gupta B, Bhattacharya A. An efficient algorithm for load partitioning in multicore shared-Tree multicast. In Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017. The International Society for Computers and Their Applications (ISCA). 2017. p. 149-153
Sinha, Koushik ; Gupta, Bidyut ; Bhattacharya, Ansuman. / An efficient algorithm for load partitioning in multicore shared-Tree multicast. Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017. The International Society for Computers and Their Applications (ISCA), 2017. pp. 149-153
@inproceedings{633c63e1de5b406c905386adef8d71b3,
title = "An efficient algorithm for load partitioning in multicore shared-Tree multicast",
abstract = "Core Based Tree (CBT) multicast involves using a single shared minimum cost tree rooted at a designated core node, to handle multiple multicasts. As the core node represents a single point of failure, recent works have explored the use of multiple candidate cores to increase the fault-Tolerance of CBT multicasting. In [20], the authors proposed utilizing these candidate cores to not only improve fault tolerance, but also for load sharing of multicast traffic between the cores. The load partitioning approach in [20] utilizes the concept of pesudo-diameter for calculating the load to be offered by a multicast source to each of the candidate cores. In this paper, we present a new multicore based load partitioning algorithm, in which each multicast source node generates a fibonaaci sequence based partitioning of its data traffic load between the candidate cores. Through simulation, we have compared the performances of our proposed Fibonacci multicore load partitioned multicast (FMLPM) algorithm with the load partitioning scheme proposed in [20]. For the mean makespan of multicast packet delay, our simulation results show that our proposed FMLPM algorithm provides on an average a 3.5{\%} improvement over the load partitioning algorithm proposed in [20]. The FMLPM algorithm has an added benefit of not requiring any floating-point operation for calculating the load to be offered to each candidate core by a multicast source node.",
keywords = "Core based tree, Fibonacci, Golden ratio, Load partitioning, Multicore multicast, Packet delay, Pseudo diameter",
author = "Koushik Sinha and Bidyut Gupta and Ansuman Bhattacharya",
year = "2017",
language = "English",
pages = "149--153",
booktitle = "Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017",
publisher = "The International Society for Computers and Their Applications (ISCA)",

}

TY - GEN

T1 - An efficient algorithm for load partitioning in multicore shared-Tree multicast

AU - Sinha, Koushik

AU - Gupta, Bidyut

AU - Bhattacharya, Ansuman

PY - 2017

Y1 - 2017

N2 - Core Based Tree (CBT) multicast involves using a single shared minimum cost tree rooted at a designated core node, to handle multiple multicasts. As the core node represents a single point of failure, recent works have explored the use of multiple candidate cores to increase the fault-Tolerance of CBT multicasting. In [20], the authors proposed utilizing these candidate cores to not only improve fault tolerance, but also for load sharing of multicast traffic between the cores. The load partitioning approach in [20] utilizes the concept of pesudo-diameter for calculating the load to be offered by a multicast source to each of the candidate cores. In this paper, we present a new multicore based load partitioning algorithm, in which each multicast source node generates a fibonaaci sequence based partitioning of its data traffic load between the candidate cores. Through simulation, we have compared the performances of our proposed Fibonacci multicore load partitioned multicast (FMLPM) algorithm with the load partitioning scheme proposed in [20]. For the mean makespan of multicast packet delay, our simulation results show that our proposed FMLPM algorithm provides on an average a 3.5% improvement over the load partitioning algorithm proposed in [20]. The FMLPM algorithm has an added benefit of not requiring any floating-point operation for calculating the load to be offered to each candidate core by a multicast source node.

AB - Core Based Tree (CBT) multicast involves using a single shared minimum cost tree rooted at a designated core node, to handle multiple multicasts. As the core node represents a single point of failure, recent works have explored the use of multiple candidate cores to increase the fault-Tolerance of CBT multicasting. In [20], the authors proposed utilizing these candidate cores to not only improve fault tolerance, but also for load sharing of multicast traffic between the cores. The load partitioning approach in [20] utilizes the concept of pesudo-diameter for calculating the load to be offered by a multicast source to each of the candidate cores. In this paper, we present a new multicore based load partitioning algorithm, in which each multicast source node generates a fibonaaci sequence based partitioning of its data traffic load between the candidate cores. Through simulation, we have compared the performances of our proposed Fibonacci multicore load partitioned multicast (FMLPM) algorithm with the load partitioning scheme proposed in [20]. For the mean makespan of multicast packet delay, our simulation results show that our proposed FMLPM algorithm provides on an average a 3.5% improvement over the load partitioning algorithm proposed in [20]. The FMLPM algorithm has an added benefit of not requiring any floating-point operation for calculating the load to be offered to each candidate core by a multicast source node.

KW - Core based tree

KW - Fibonacci

KW - Golden ratio

KW - Load partitioning

KW - Multicore multicast

KW - Packet delay

KW - Pseudo diameter

UR - http://www.scopus.com/inward/record.url?scp=85016594193&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85016594193&partnerID=8YFLogxK

M3 - Conference contribution

SP - 149

EP - 153

BT - Proceedings of the 32nd International Conference on Computers and Their Applications, CATA 2017

PB - The International Society for Computers and Their Applications (ISCA)

ER -