Dr. multicast
Published 13 April 2010
Ýmir Vigfússon, Hussam Abu-Libdeh, Mahesh Balakrishnan, Ken Birman, Robert Burgess, Gregory Chockler
Citations77
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
Experimental evaluation shows that MCMD allows a large number of IPMC groups to be used without disruption, restoring a powerful group communication primitive to its traditional role.
Abstract
IP Multicast (IPMC) in data centers becomes disruptive when the technology is used by a large number of groups, a capability desired by event notification systems. We trace the problem to root causes, and introduce Dr. Multicast (MCMD), a system that eliminates the issue by mapping IPMC operations to a combination of point-to-point unicast and traditional IPMC transmissions guaranteed to be safe. MCMD optimizes the use of IPMC addresses within a data center by merging similar multicast groups in a principled fashion, while simultaneously respecting hardware limits expressed through administrator-controlled policies. The system is fully transparent, making it backward-compatible with commodity hardware and software found in modern data centers. Experimental evaluation shows that MCMD allows a large number of IPMC groups to be used without disruption, restoring a powerful group communication primitive to its traditional role.
Keywords
Computer Science
OPAL (Open@LaTrobe) (La Trobe University)Power-law distributions in empirical data
6,724 Citations2018Aaron Clauset, Cosma Rohilla Shalizi +1 more
Group formation in large social networks
1,820 Citations2006Lars Bäckström, D.P. Huttenlocher +2 more
It is found that the propensity of individuals to join communities, and of communities to grow rapidly, depends in subtle ways on the underlying network structure, and decision-tree techniques are used to identify the most significant structural determinants of these properties.
Host extensions for IP multicasting
1,340 Citations1989S. Deering
ACM Transactions on Computer SystemsMulticast routing in datagram internetworks and extended LANs
1,179 Citations1990S. Deering, David R. Cheriton
Extensions to two common internetwork routing algorithms are specified to support low-delay datagram multicasting beyond a single LAN and how the use of multicast scope control and hierarchical multicast routing allows the multicast service to scale up to large internetworks is discussed.
IEEE/ACM Transactions on NetworkingA reliable multicast framework for light-weight sessions and application level framing
1,165 Citations1997Sally Floyd, Van Jacobson +3 more
IEEE NetworkDeployment issues for the IP multicast service and architecture
787 Citations2000Christophe Diot, Brian Neil Levine +3 more
This work examines the issues that have limited the commercial deployment of IP multicast from the viewpoint of carriers, and analyzes where the model fails and what it does not offer, and discusses requirements for successful deployment of multicast services.
Multicast security: a taxonomy and some efficient constructions
700 Citations1999Ran Canetti, Juan A. Garay +4 more
A taxonomy of multicast scenarios on the Internet and an improved solution to the key revocation problem are presented, which can be regarded as a 'midpoint' between traditional message authentication codes and digital signatures.
ACM Transactions on Computer SystemsBimodal multicast
652 Citations1999Ken Birman, Mark Hayden +4 more
This article introduces the protocol, provides a theoretical analysis of its behavior, review experimental results, and discusses some candidate applications, confirming that bimodal multicast is reliable, scalable, and that the protocol provides remarkably stable delivery throughput.
Feedback effects between similarity and social influence in online communities
603 Citations2008David Crandall, Dan Cosley +3 more
Clear feedback effects between the two factors are found, with rising similarity between two individuals serving, in aggregate, as an indicator of future interaction -- but with similarity then continuing to increase steadily, although at a slower rate, for long periods after initial interactions.
A Gossip-Style Failure Detection Service
500 Citations1998Robbert van Renesse, Yaron Minsky +1 more
A new protocol based on gossiping is described that does scale well and provides timely detection, and is extended to discover and leverage the underlying network topology for much improved resource utilization.
LIPSIN
434 Citations2009Petri Jokela, András Zahemszky +3 more
A novel multicast forwarding fabric suitable for large-scale topic-based publish/subscribe is proposed, and due to very simple forwarding decisions and small forwarding tables, the fabric may be more energy efficient than the currently used ones.
The dynamics of viral marketing
396 Citations2006Jure Leskovec, Lada A. Adamic +1 more
A model is presented that successfully identifies product and pricing categories for which viral marketing seems to be very effective and how the recommendation network grows over time and how effective it is from the viewpoint of the sender and receiver of the recommendations is established.
Extending equation-based congestion control to multicast applications
246 Citations2001Joerg Widmer, Mark Handley
This paper improves upon the well-known approach of using exponentially weighted random timers by biasing feedback in favor of low-rate receivers while still preventing a response implosion in TFMCC.
SIAM Journal on ComputingMinimal NFA Problems are Hard
246 Citations1993Tao Jiang, Bala Ravikumar
This work studies the complexity of decision problems for finite automata and presents many fundamental decision problems which are computationally intractable even when the input is a DFA or a NFA with limited nondeterminism.
An Overview of Source-Specific Multicast (SSM)
226 Citations2003Supratik Bhattacharyya, H Holbrook +21 more
How the SSM service model addresses the challenges faced in inter-domain multicast deployment, changes needed to routing protocols and applications to deploy SSM and interoperability issues with current multicast service models are discussed.
IEEE NetworkA survey of security issues in multicast communications
196 Citations1999M.J. Moyer, J. R. Rao +1 more
This survey discusses issues in a group communication architecture, group key management, and message source authentication, and proposed solutions to them in secure multicast.
Redundancy in network traffic
193 Citations2009Ashok Anand, C. R. Muthukrishnan +2 more
A new redundancy elimination algorithm based on Winnowing is presented that outperforms the widely-used Rabin fingerprint-based algorithm by 5-10% on most traces and by as much as 35% in some traces, implying that pushing redundancy elimination capability to the end hosts could obtain most of the middlebox's bandwidth savings.
ACM SIGCOMM Computer Communication ReviewPacket caches on routers
186 Citations2008Ashok Anand, Archit Gupta +3 more
ACM SIGACT NewsToward a cloud computing research agenda
186 Citations2009Ken Birman, Gregory Chockler +1 more
The 2008 LADIS workshop on Large Scale Distributed Systems brought together leaders from the commercial cloud computing community with researchers working on a variety of topics in distributed computing, and yielded some surprises.
IEEE NetworkSecurity issues and solutions in multicast content distribution: a survey
146 Citations2003Paul Judge, Mostafa Ammar
Client behavior and feed characteristics of RSS, a publish-subscribe system for web micronews
139 Citations2005Hongzhou Liu, Venugopalan Ramasubramanian +1 more
Packet caches on routers
116 Citations2008Ashok Anand, Archit Gupta +3 more
It is argued that far more significant network-wide benefits can be derived by redesigning network routing protocols to leverage the universal deployment of packet-level redundant content elimination as a universal primitive on all Internet routers.
Multicast-specific security threats and counter-measures
109 Citations2002Tony Ballardie, Jon Crowcroft
This work proposes a version of the IGMP protocol that can reliably enforce subnet-level group access control and describes a scalable mechanism to control multicast traffic in transit that can, for example, prevent a misbehaving source from causing undue congestion over the wide-area.
Aggregated multicast: an approach to reduce multicast state
100 Citations2002Aiguo Fei, Jun‐Hong Cui +2 more
This paper proposes an approach to reduce the multicast forwarding state, in which multiple groups are forced to share a single delivery tree and initial simulation results suggest that this method can reduce multicast state significantly.
An analysis of multicast forwarding state scalability
65 Citations2002Tina Wong, Randy H. Katz
It is found that the increase in peering among Internet backbone networks has led to more multicast forwarding state at a handful of core domains, but less state in the rest of the domains.
Gothic: a group access control architecture for secure multicast and anycast
64 Citations2003Paul Judge, Mostafa Ammar
Gothic is proposed, a complete architecture for providing group access control for multicast and anycast that improves scalability over previous solutions while maintaining or increasing the level of security.
An evaluation of preference clustering in large-scale multicast applications
63 Citations2002Tina Wong, Randy H. Katz +1 more
This work presents a clustering framework with a two-phase algorithm: a bootstrapping phase that groups new sources and receivers together, and an adaptation phase that re-groups them in reaction to changes, found that clustering successfully exploits preference similarity and utilizes network resources more efficiently than when it is not used.
Channelization problem in large scale data dissemination
58 Citations2005Micah Adler, Zihui Ge +3 more
This work formally defines two versions of the channelization problem and subscription problem, and finds that the subscription problem is NP-complete when one flow can be assigned to multiple multicast groups.
Ricochet: lateral error correction for time-critical multicast
34 Citations2007Mahesh Balakrishnan, Ken Birman +2 more
Ricochet is evaluated on a 64-node cluster with up to 1024 groups per node: under various loss rates, it recovers almost all packets using LEC in tens of milliseconds and the remainder with reactive traffic within 200 milliseconds.
eCommons (Cornell University)Dr. Multicast: Rx for Datacenter Communication Scalability.
23 Citations2008Ýmir Vigfússon, Hussam Abu-Libdeh +3 more
Dr. Multicast
23 Citations2008Ýmir Vigfússon, Hussam Abu-Libdeh +3 more
This work introduces Dr. Multicast (MCMD), a system that maps IPMC operations to a combination of point-to-point unicast and traditional IPMC transmissions and argues that with the resulting range of options, IPMC no longer represents a threat and can therefore be used much more widely.
High throughput reliable message dissemination
22 Citations2004Boaz Carmeli, Gidon Gershinsky +5 more
It is shown that in typical real-world messaging applications, performing message aggregation can increase throughput by order of magnitude and is indicated that the achievable throughput of messaging applications is determined by the number of packets- per-second, rather than bytes-per-second the receiver or a transmitter should handle.
ACM SIGOPS Operating Systems ReviewBulletin board
11 Citations2010Vita Bortnikov, Gregory Chockler +2 more
The design and early experience with a completely new implementation of the Bulletin Board, a topicbased distributed shared memory service employed by commercial-grade application middleware, is presented to achieve robustness and administrative simplicity with adequate latency and costs at the required throughput and scale.
eCommons (Cornell University)Affinity In Distributed Systems
6 Citations2010Ýmir Vigfússon
A transparent and backward-compatible layer called Dr. Mul-ticast is proposed to allow data center administrators to enable IPMC for large numbers of groups without causing stability issues and a platform called GO (for Gossip Objects) is proposed that bounds the node's bandwidth use to a customizable limit, prohibiting applications from joining groups that would cause the limit to be exceeded.