Download presentation
Presentation is loading. Please wait.
Published byGriselda Melton Modified over 9 years ago
1
PAGE: A Partition Aware Graph Computation Engine Yingxia Shao, Junjie Yao, Bin Cui, Lin Ma EECS, Peking University, China
2
Agenda Background Design of PAGE Experiment result Conclusion 2/19
3
Background Prevalent large scale graphs – Social networks – Web graph – … Graph computing systems – Pregel (Google) – Giraph (Apache) – GPS (Stanford) – GraphLab (CMU) – … 3/19
4
Background Graph Partitioning – Offline approach METIS (Karypis Lab) – Online approach Streaming partitioning Linear Deterministic Greedy(LDG) algorithm (I. Stanton) 4/19 Problem: The existing graph computation systems cannot efficiently integrate the high-quality graph partitioning.
5
Inefficient partition integrating 5/19 The high-quality graph partitioning leads to the worse overall performance. The graph partitioning quality is improved from left to right. Running PageRank on Giraph with six different graph partition qualities.
6
Motivation of the PAGE Call for a novel graph computation engine to efficiently integrate graph partitioning with various qualities. 6/19
7
Agenda Background Design of PAGE Experiment result Conclusion 7/19
8
Message processor 8/19
9
Inefficient partition integrating 9/19 The local message processing cost dominates the overall cost. The existing systems cannot provide enough local message processor. Running PageRank on Giraph with six different graph partition qualities.
10
Overview of the PAGE PAGE applies adaptively tuning mechanism and new cooperation methods. 10/19
11
New Designed PAGE Worker 11/19
12
Dual Concurrent Message Processor First type concurrency – A remote MP and a local MP are embedded Second type concurrency – A set of message process units are contained by each message processor The concurrency is automatically determined by the system itself. 12/19
13
Dynamic Concurrency Control Model The DCCM determines the proper parameters, such as nmp, nmp l, nmp r. The DCCM is built on top of two heuristic rules. – Ability Lower-bound. – Workload Balance Ratio. Monitor – Tracks the necessary metrics 13/19
14
Agenda Background Design of PAGE Experiment result Conclusion 14/19
15
Environment & Datasets Experiment Environment – a 24 nodes cluster Dataset: the uk-2007-05-u. – Undirected – Vertex #: 105,153,952 – Edge #: 6,603,753,128 Benchmark: PageRank SchemeEdge Cut Random98.52% LDG182.88% LDG275.69% LDG366.37% LDG456.34% METIS3.48% Partition qualities 15/19 Balance factor: < 1%.
16
Partition Awareness in PAGE PAGEGiraph 16/19
17
Compare with the naive solution 17/19 * The Giraph-GPSop is the naive solution.
18
Contribution & Conclusion We identify the problem of partition unaware inefficiency. We set up a new partition aware graph computation engine, PAGE. We design a Dynamic Concurrency Control Model based on several heuristic rules to better profile the characters of graph partition. At last, we demonstrate PAGE’s robustness and efficiency on different graph partition qualities. 18/19
19
19/19 Email: simon0227@gmail.com
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.