PEGASUS: A Peta-Scale Graph Mining System Implementation and Observations

By U. Kang, Charalampos E. Tsourakakis, Christos Faloutsos. 2009 Ninth IEEE International Conference on Data Mining, 2009

733 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Information Systems, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/337bb008eb531c4e8152d6c99423e715ce0d9b05

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Information Systems

In Statistical and Nonlinear Physics

Citations in the corpus

Cyclades: Conflict-free Asynchronous Machine Learning (2016, methodology): “In that case, we can use a very simple “push-label” CC algorithm, whose cost on P cores and arbitrary graphs with E edges and max degree ∆ is O (max { , ∆ } · C max ) , where C max is the size of the longest-shortest path, or the diameter of the graph [KTF09].”
MACH: Fast Randomized Tensor Decompositions (2009, methodology): “As future work, we plan to implement tensor decomposition algorithms using Hadoop, the open source version of MapReduce [7] for the PEGASUS project [20]”
Pregel: a system for large-scale graph processing (2010, background): “Graph algorithms can be written as a series of chained MapReduce invocations [11, 30].”
Book Reviews: Data-Intensive Text Processing with MapReduce by Jimmy Lin and Chris Dyer (2010, methodology): “[80] presented an approach to estimating the diameter of large graphs using MapReduce and a library for graph mining [81]; Cohen [39] discussed”
Mizan: a system for dynamic load balancing in large-scale graph processing (2013, background): “, Pregel [23], HADI [14], PEGASUS [13] and X-RIME [31]).”
Graphicionado: A high-performance and energy-efficient accelerator for graph analytics (2016, background): “This renewed interest spurred the software community to develop more efficient graph analytics processing frameworks [4, 5, 23, 25, 33, 35, 46] as well as the hardware community to create hardware that can execute graph analytics applications with more efficiency than what the off-the-shelf, general”
Handbook of Big Data (2016, background): “The importance of this primitive is frequently suggested in the literature for scalable graph algorithms [18,35].”
GraphLab: A Distributed Framework for Machine Learning in the Cloud (2011, methodology): “We chose to compare against Hadoop, due to its wide acceptance in the Machine Learning community for large scale computation (for example the Mahout project [2, 45]).”
GPS: a graph processing system (2013, methodology): “HCC [22] is an algorithm to find the weakly connected components of an undirected graph: First, every vertex sets its value to its own ID.”
Distributed GraphLab: A Framework for Machine Learning in the Cloud (2012, methodology): “Pegasus [20] is a collection of algorithms for mining large graphs using Hadoop.”
SNAP: A General Purpose Network Analysis and Graph Mining Library (2016, background): “Examples of such systems include Pregel [Malewicz et al. 2010], PowerGraph [Gonzalez et al. 2012], Pegasus [Kang et al. 2009], and GraphX [Xin et al. 2013].”
GigaTensor: scaling tensor analysis up by 100 times - algorithms and discoveries (2012, methodology): “Due to its excellent scalability, ease of use, cost advantage, HADOOP has been used for important graph mining algorithms (see [25, 16]).”
Fast personalized PageRank on MapReduce (2011, background): “Its effectiveness and simplicity has resulted in its implementation by different internet companies [9, 13, 5, 22], and widespread adoption for a wide range of applications [19], including large scale graph computations [15, 16 ].,MapReduce provides an excellent tool for large scale graph processin”
Distributed GraphLab: A Framework for Machine Learning in the Cloud (2012, methodology): “port for basic graph computation. In contrast, GraphLab focuses on iterative graph structured computation. There are several notable projects focused on using MapReduce for graph computation. Pegasus [20] isa collectionof algorithms for mining large graphs using Hadoop. Surfer [7] extends MapReduce ”
MapReduce for Parallel Reinforcement Learning (2011, background): “See more discussions about designing matrix operations in [8, 11].”
GraphMat: High performance graph analytics made productive (2015, methodology): “PEGASUS is based on Map-Reduce and suffers from poor performance due to I/O bottlenecks compared to in-memory frameworks.”