Piggybacking on social networks

Aristides Gionis, Flavio Junqueira, Vincent Leroy, Marco Serafini, Ingmar Weber

Research output: Chapter in Book/Report/Conference proceedingChapter

12 Citations (Scopus)

Abstract

The popularity of social-networking sites has increased rapidly over the last decade. A basic functionalities of social-networking sites is to present users with streams of events shared by their friends. At a systems level, materialized per-user views are a common way to assemble and deliver such event streams on-line and with low latency. Access to the data stores, which keep the user views, is a major bottleneck of social-networking systems. We propose to improve the throughput of these systems by using social piggybacking, which consists of processing the requests of two friends by querying and updating the view of a third common friend. By using one such hub view, the system can serve requests of the first friend without querying or updating the view of the second. We show that, given a social graph, social piggybacking can minimize the overall number of requests, but computing the optimal set of hubs is an NP-hard problem. We propose an θ(log n) approximation algorithm and a heuristic to solve the problem, and evaluate them using the full Twitter and Flickr social graphs, which have up to billions of edges. Compared to existing approaches, using social piggybacking results in similar throughput in systems with few servers, but enables substantial throughput improvements as the size of the system grows, reaching up to a 2-factor increase. We also evaluate our algorithms on a real social networking system prototype and we show that the actual increase in throughput corresponds nicely to the gain anticipated by our cost function.

Original languageEnglish
Title of host publicationProceedings of the VLDB Endowment
Pages409-420
Number of pages12
Volume6
Edition6
Publication statusPublished - 2013

Fingerprint

Throughput
Approximation algorithms
Cost functions
Computational complexity
Servers
Processing

ASJC Scopus subject areas

  • Computer Science (miscellaneous)
  • Computer Science(all)

Cite this

Gionis, A., Junqueira, F., Leroy, V., Serafini, M., & Weber, I. (2013). Piggybacking on social networks. In Proceedings of the VLDB Endowment (6 ed., Vol. 6, pp. 409-420)

Piggybacking on social networks. / Gionis, Aristides; Junqueira, Flavio; Leroy, Vincent; Serafini, Marco; Weber, Ingmar.

Proceedings of the VLDB Endowment. Vol. 6 6. ed. 2013. p. 409-420.

Research output: Chapter in Book/Report/Conference proceedingChapter

Gionis, A, Junqueira, F, Leroy, V, Serafini, M & Weber, I 2013, Piggybacking on social networks. in Proceedings of the VLDB Endowment. 6 edn, vol. 6, pp. 409-420.
Gionis A, Junqueira F, Leroy V, Serafini M, Weber I. Piggybacking on social networks. In Proceedings of the VLDB Endowment. 6 ed. Vol. 6. 2013. p. 409-420
Gionis, Aristides ; Junqueira, Flavio ; Leroy, Vincent ; Serafini, Marco ; Weber, Ingmar. / Piggybacking on social networks. Proceedings of the VLDB Endowment. Vol. 6 6. ed. 2013. pp. 409-420
@inbook{354ed6344a784813b887ea6a9441594d,
title = "Piggybacking on social networks",
abstract = "The popularity of social-networking sites has increased rapidly over the last decade. A basic functionalities of social-networking sites is to present users with streams of events shared by their friends. At a systems level, materialized per-user views are a common way to assemble and deliver such event streams on-line and with low latency. Access to the data stores, which keep the user views, is a major bottleneck of social-networking systems. We propose to improve the throughput of these systems by using social piggybacking, which consists of processing the requests of two friends by querying and updating the view of a third common friend. By using one such hub view, the system can serve requests of the first friend without querying or updating the view of the second. We show that, given a social graph, social piggybacking can minimize the overall number of requests, but computing the optimal set of hubs is an NP-hard problem. We propose an θ(log n) approximation algorithm and a heuristic to solve the problem, and evaluate them using the full Twitter and Flickr social graphs, which have up to billions of edges. Compared to existing approaches, using social piggybacking results in similar throughput in systems with few servers, but enables substantial throughput improvements as the size of the system grows, reaching up to a 2-factor increase. We also evaluate our algorithms on a real social networking system prototype and we show that the actual increase in throughput corresponds nicely to the gain anticipated by our cost function.",
author = "Aristides Gionis and Flavio Junqueira and Vincent Leroy and Marco Serafini and Ingmar Weber",
year = "2013",
language = "English",
volume = "6",
pages = "409--420",
booktitle = "Proceedings of the VLDB Endowment",
edition = "6",

}

TY - CHAP

T1 - Piggybacking on social networks

AU - Gionis, Aristides

AU - Junqueira, Flavio

AU - Leroy, Vincent

AU - Serafini, Marco

AU - Weber, Ingmar

PY - 2013

Y1 - 2013

N2 - The popularity of social-networking sites has increased rapidly over the last decade. A basic functionalities of social-networking sites is to present users with streams of events shared by their friends. At a systems level, materialized per-user views are a common way to assemble and deliver such event streams on-line and with low latency. Access to the data stores, which keep the user views, is a major bottleneck of social-networking systems. We propose to improve the throughput of these systems by using social piggybacking, which consists of processing the requests of two friends by querying and updating the view of a third common friend. By using one such hub view, the system can serve requests of the first friend without querying or updating the view of the second. We show that, given a social graph, social piggybacking can minimize the overall number of requests, but computing the optimal set of hubs is an NP-hard problem. We propose an θ(log n) approximation algorithm and a heuristic to solve the problem, and evaluate them using the full Twitter and Flickr social graphs, which have up to billions of edges. Compared to existing approaches, using social piggybacking results in similar throughput in systems with few servers, but enables substantial throughput improvements as the size of the system grows, reaching up to a 2-factor increase. We also evaluate our algorithms on a real social networking system prototype and we show that the actual increase in throughput corresponds nicely to the gain anticipated by our cost function.

AB - The popularity of social-networking sites has increased rapidly over the last decade. A basic functionalities of social-networking sites is to present users with streams of events shared by their friends. At a systems level, materialized per-user views are a common way to assemble and deliver such event streams on-line and with low latency. Access to the data stores, which keep the user views, is a major bottleneck of social-networking systems. We propose to improve the throughput of these systems by using social piggybacking, which consists of processing the requests of two friends by querying and updating the view of a third common friend. By using one such hub view, the system can serve requests of the first friend without querying or updating the view of the second. We show that, given a social graph, social piggybacking can minimize the overall number of requests, but computing the optimal set of hubs is an NP-hard problem. We propose an θ(log n) approximation algorithm and a heuristic to solve the problem, and evaluate them using the full Twitter and Flickr social graphs, which have up to billions of edges. Compared to existing approaches, using social piggybacking results in similar throughput in systems with few servers, but enables substantial throughput improvements as the size of the system grows, reaching up to a 2-factor increase. We also evaluate our algorithms on a real social networking system prototype and we show that the actual increase in throughput corresponds nicely to the gain anticipated by our cost function.

UR - http://www.scopus.com/inward/record.url?scp=84881236172&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84881236172&partnerID=8YFLogxK

M3 - Chapter

VL - 6

SP - 409

EP - 420

BT - Proceedings of the VLDB Endowment

ER -