K-Core Decomposition of Large Networks on a Single PC

  title={K-Core Decomposition of Large Networks on a Single PC},
  author={Wissam Khaouid and Marina Barsky and Venkatesh Srinivasan and Alex Thomo},
  journal={Proc. VLDB Endow.},
Studying the topology of a network is critical to inferring underlying dynamics such as tolerance to failure, group behavior and spreading patterns. k-core decomposition is a well-established metric which partitions a graph into layers from external to more central vertices. In this paper we aim to explore whether k-core decomposition of large networks can be computed using a consumer-grade PC. We feature implementations of the "vertex-centric" distributed protocol introduced by Montresor, De… 

Figures and Tables from this paper

Computation of K-Core Decomposition on Giraph
By analyzing the results, it is concluded that Giraph is faster than GraphChi when dealing with large data, however, since worker nodes need time to communicate with each other, Giraph was not very efficient for small data.
Efficient Computation of Probabilistic Core Decomposition at Web-Scale
A peeling algorithm is proposed to compute the core decomposition of a probabilistic graph that scales to very large graphs and is orders of magnitude faster than the state-of-the-art approach.
Parallel k-Core Decomposition on Multicore Platforms
  • H. Kabir, Kamesh Madduri
  • Computer Science, Mathematics
    2017 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)
  • 2017
This work presents a new shared-memory parallel algorithm called PKC for k-core decomposition on multicore platforms by reducing synchronization overhead and creating a smaller graph to process high-degree vertices and shows that PKC consistently outperforms implementations of other methods on a 32-core multicore server and on a collection of large sparse graphs.
Parallel Batch-Dynamic k-Core Decomposition
This work presents the first approximate static k-core algorithm with linear work and polylogarithmic depth, and shows that on a 30-core machine with two-way hyper-threading, the implementation achieves up to a 3.9x speedup in the static case over the previous state-of-the-art parallel algorithm.
K-Truss Decomposition of Large Networks on a Single Consumer-Grade Machine
With their optimized implementation, this work shows that it can efficiently compute k-truss decomposition of large networks (e.g., a graph with 1.2 billion edges) on a single consumer-grade machine.
Efficient Computation of Importance Based Communities in Web-Scale Networks Using a Single Machine
The goal is to scale-up the computation of top-r, k-core communities to web-scale graphs of tens of billions of edges using a single consumer-level machine within a reasonable amount of time.
The core decomposition of networks: theory, algorithms and applications
In this survey, an in-depth discussion of core decomposition is performed, focusing mainly on the basic theory and fundamental concepts, the algorithmic techniques proposed for computing it efficiently under different settings, and the applications that can benefit significantly from it.
Hierarchical Core Maintenance on Large Dynamic Graphs
In this paper, in-depth analyses on the structural properties of the hierarchy are conducted, and well-designed local update techniques are proposed that significantly outperform the baselines on runtime by up to 3 orders of magnitude.
K-core decomposition on super large graphs with limited resources
The experimental results show that the consumption of resources can be significantly reduced, the calculation on large-scale graphs becomes more stable than the existing methods, and the divide-and-conquer technique is proposed.
I/O efficient Core Graph Decomposition at web scale
This paper proposes a semi-external algorithm and two optimized algorithms for I/O efficient core decomposition using very simple structures and data access model, and is the first to handle a web graph with 978.5 million nodes and 42.6 billion edges using less than 4.2 GB memory.


Efficient core decomposition in massive networks
This paper proposes the first external-memory algorithm for core decomposition in massive graphs and demonstrates the efficiency of the algorithm on real networks with up to 52.9 million vertices and 1.65 billion edges.
Pregel: a system for large-scale graph processing
A model for processing large graphs that has been designed for efficient, scalable and fault-tolerant implementation on clusters of thousands of commodity computers, and its implied synchronicity makes reasoning about programs easier.
GraphChi: Large-Scale Graph Computation on Just a PC
This work presents GraphChi, a disk-based system for computing efficiently on graphs with billions of edges, and builds on the basis of Parallel Sliding Windows to propose a new data structure Partitioned Adjacency Lists, which is used to design an online graph database graphChi-DB.
K-core decomposition of Internet graphs: hierarchies, self-similarity and measurement biases
It is found that the k-core analysis provides an interesting characterization of the fluctuations and incompleteness of maps as well as information helping to discriminate the original underlying structure.
Core decomposition of uncertain graphs
It is shown that core decomposition of uncertain graphs can be carried out efficiently as well, and the definitions and methods are evaluated on a number of real-world datasets and applications, such as influence maximization and task-driven team formation.
Community detection in graphs
Characterization of Graphs Using Degree Cores
This work introduces the concept of a hierarchical degree core tree as a novel way of summarizing the structure of massive graphs and extracts features related to the graph's local structure from these hierarchical trees.
K-core Organization of Complex Networks
It is shown that in networks with a finite mean number zeta2 of the second-nearest neighbors, the emergence of a k-core is a hybrid phase transition, and in contrast, ifZeta2 diverges, the networks contain an infinite sequence of k-cores which are ultrarobust against random damage.
Distributed GraphLab : A Framework for Machine Learning and Data Mining in the Cloud
This paper develops graph based extensions to pipelined locking and data versioning to reduce network congestion and mitigate the effect of network latency, and introduces fault tolerance to the GraphLab abstraction using the classic Chandy-Lamport snapshot algorithm.
Preventing Unraveling in Social Networks: The Anchored k-Core Problem
A model of user engagement in social networks, where each player incurs a cost to remain engaged but derives a benefit proportional to the number of engaged neighbors, is considered, and strong inapproximability results for general graphs and k≥3 are proved.