Implementation of GPU based CCN Router

Slides:



Advertisements
Similar presentations
IP Router Architectures. Outline Basic IP Router Functionalities IP Router Architectures.
Advertisements

Multimedia and Mobile communications Laboratory CCN 1 DK Han Junghwan Song Computer Networks Practice.
M. Waldvogel, G. Varghese, J. Turner, B. Plattner Presenter: Shulin You UNIVERSITY OF MASSACHUSETTS, AMHERST – Department of Electrical and Computer Engineering.
IP Routing Lookups Scalable High Speed IP Routing Lookups.
KARL NADEN – NETWORKS (18-744) FALL 2010 Overview of Research in Router Design.
1 Architectural Results in the Optical Router Project Da Chuang, Isaac Keslassy, Nick McKeown High Performance Networking Group
Chapter 9 Classification And Forwarding. Outline.
(part 3).  Switches, also known as switching hubs, have become an increasingly important part of our networking today, because when working with hubs,
Proxy-assisted Content Sharing Using Content Centric Networking (CCN) for Resource-limited Mobile Consumer Devices Jihoon Lee, Dae Youb Kim IEEE Transactions.
WINLAB Comparing Alternative Approaches for Networking of Named Objects in the Future Internet Akash Baid, Tam Vu, Dipankar Raychaudhuri WINLAB, Rutgers.
ECE 526 – Network Processing Systems Design Network Processor Architecture and Scalability Chapter 13,14: D. E. Comer.
Lecturer: Ghadah Aldehim
Application-Layer Anycasting By Samarat Bhattacharjee et al. Presented by Matt Miller September 30, 2002.
A 50-Gb/s IP Router 참고논문 : Craig Partridge et al. [ IEEE/ACM ToN, June 1998 ]
CS848 Class Project: A Survey on QoS for Multi-tier Web Systems Huaning(Mike) Nie
Wire Speed Packet Classification Without TCAMs ACM SIGMETRICS 2007 Qunfeng Dong (University of Wisconsin-Madison) Suman Banerjee (University of Wisconsin-Madison)
CUDA Optimizations Sathish Vadhiyar Parallel Programming.
Network-Coding Multicast Networks With QoS Guarantees Yuanzhe Xuan and Chin-Tau Lea, Senior Member, IEEE IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 19,
NUOL Internet Application Services Midterm presentation 22 nd March, 2004.
1 UNIT 13 The World Wide Web Lecturer: Kholood Baselm.
Scalable High Speed IP Routing Lookups Scalable High Speed IP Routing Lookups Authors: M. Waldvogel, G. Varghese, J. Turner, B. Plattner Presenter: Zhqi.
Efficient Cache Structures of IP Routers to Provide Policy-Based Services Graduate School of Engineering Osaka City University
OSPF Open Shortest Path First. Table of Content  IP Routes  OSPF History  OSPF Design  OSPF Link State  OSPF Routing Table  OSPF Data Packets.
Capstone Project Fall Course Information Instructor Ye Zhao –Office: MSB 220 – Fall 2015 (MSB162) –Time: Tue, Thu 10:45am.
COS 420 Day 15. Agenda Finish Individualized Project Presentations on Thrusday Have Grading sheets to me by Friday Group Project Discussion Goals & Timelines.
Modularized Two-Stages Vertical Handoff Scheme in Integration of WWAN and WLAN Shimin Li Ying Wang Presented by Shimin Li E&CE750.
CS 6401 Overlay Networks Outline Overlay networks overview Routing overlays Resilient Overlay Networks Content Distribution Networks.
Sunpyo Hong, Hyesoon Kim
: MobileIP. : r Goal: Allow machines to roam around and maintain IP connectivity r Problem: IP addresses => location m This is important for efficient.
1 UNIT 13 The World Wide Web. Introduction 2 The World Wide Web: ▫ Commonly referred to as WWW or the Web. ▫ Is a service on the Internet. It consists.
Mobile IP THE 12 TH MEETING. Mobile IP  Incorporation of mobile users in the network.  Cellular system (e.g., GSM) started with mobility in mind. 
Cost-Effective Video Streaming Techniques Kien A. Hua School of EE & Computer Science University of Central Florida Orlando, FL U.S.A.
Exploiting Graphics Processors for High-performance IP Lookup in Software Routers Jin Zhao, Xinya Zhang, Xin Wang, Yangdong Deng, Xiaoming Fu IEEE INFOCOM.
Shengling Wang; Yong Cui; Das, S.; Mingwei Xu; Communications Workshops, ICC Workshops '08. IEEE International Conference on May 19-23, 2008 Page(s):441.
Optimising Streaming Systems with SDN/P4/NetFPGA
Auction-based in-network caching in Information-centric networks Workshop ACROSS, 16th of September 2016 | Lucia D’Acunto.
University of Maryland College Park
HTTP AND ABSTRACTION ON THE INTERNET
Cellular IP: A New Approach to Internet Host Mobility
Whirlwind Tour Of Lectures So Far
Section 4 – Computer Networks
Multicast Outline Multicast Introduction and Motivation DVRMP.
NDN (Named Data Networking)
Vocabulary Prototype: A preliminary sketch of an idea or model for something new. It’s the original drawing from which something real might be built or.
What Are Routers? Routers are an intermediate system at the network layer that is used to connect networks together based on a common network layer protocol.
Addressing: Router Design
Vocabulary Prototype: A preliminary sketch of an idea or model for something new. It’s the original drawing from which something real might be built or.
Lecture 5: GPU Compute Architecture
© 2002, Cisco Systems, Inc. All rights reserved.
Q and A, Ch. 21 CS332, Fall 2017 Victor Norman.
Chapter 9: Multiarea OSPF
Statistical Optimal Hash-based Longest Prefix Match
Foundations of Networking
Introduction There are many situations in which we might use replicated data Let’s look at another, different one And design a system to work well in that.
Modelling the internet
2 - IP Routing.
Edge computing (1) Content Distribution Networks
Static Routing 1st semester
Modelling the internet
Chapter 9: Multiarea OSPF
Net 323 D: Networks Protocols
EE 122: Lecture 7 Ion Stoica September 18, 2001.
Chapter 9: Multiarea OSPF
Route Metric Proposal Date: Authors: July 2007 Month Year
EE 122: Lecture 22 (Overlay Networks)
Lecture 18: Coherence and Synchronization
Author: Xianghui Hu, Xinan Tang, Bei Hua Lecturer: Bo Xu
COSC 3330/6308 Materials on the Final
Requirements of Computing in Network
Mobile IP Neil Tang 11/12/2008 CS440 Computer Networks.
Presentation transcript:

Implementation of GPU based CCN Router [EE817] Project Proposal May 18, 2017 Implementation of GPU based CCN Router [Group 7] 20164405 Junhee Lee 20175053 Seyeon Kim

Contents Background Motivation Problem Description How To Solve Task Allocation for Each Member Schedule References

Conventional IP Networks Background Conventional IP Networks What is IP Networking ? Routing based on IP address Focus on end-to-end communication Router : Path determination based on ip-routing table What is the conventional IP networking? It routes the packets based on IP address. And, It focuses on end-to-end communication. Router determines path based on ip routing table. The packet forwarding information is in the routing table.

Content Centric Networking(CCN) Background Content Centric Networking(CCN) CCN Routing : Instead of IP address, use contents name Caching : Routers cache some contents like little server Why CCN? Rapidly increasing demands on contents Need to meet low latency, to use links more efficiently CCN is an abbreviation of content centric networking. Before explain what is CCN, I’m gonna firstly explain why CCN is necessary. These days, the demanding of ‘contents’ like video or audio.. increases rapidly. You can see in this picture. So, for example, youtube server may suffer from lots of traffic requests. It degrades networking performance. In this case, ‘where is the information’ is not important, the contents itself becomes important. Therefore, the CCN methodology is devised. It uses contents’ name instead of IP address for routing. And, CCN routers cache the contents to avoid bottleneck phenomenon. I’ll explain in details on the next slide. [1] Ericsson Mobility Report , June 2016

Background CCN Routing Process CCN Routing CCN Routers caches some contents CCN Routers manages contents based routing table Clients asks some contents to the server, and the contents is passed back to client through CCN routers. During passing, the CCN routers store the contents for a certain time. And when another client asks the same contents, the router passes the stored contents directly without connecting server.

Motivation What is Research Issue ? CCN routing table Backbone CCN routers have large-scale forwarding tables compared to the current IP routing tables Contents names are far more complex than IP address(e.g. even hundreds of characters) Fast look-up is research challenge [2] Updating(publish/delete) routing table occurs frequently Need to consider efficient method on memory occupation, throughput, latency and fast incremental update Backbone CCN routers which is top hierarchy of routers have large-scale forwarding tables compared with the conventional IP routing tables Contents names are far more complex that IP address(e.g. even hundreds of characters). So, look-up the table fast is challenging. Updating(publish/delete) routing table occurs frequently And for the fast table look-up, we need to consider efficient method on memory occupation, throughput, latency and fast incremental update.

Problem Description Fast CCN routing table look-up Whenever a content is requested, CCN router will search the name in a routing table This searching process may be boosted through parallel programing using GPU Fast CCN routing table update Fast updating routing table is also important This updating(publish/delete) may be boosted through parallel programming using GPU The problem is, first, fast CCN routing table look-up method is required. Whenever a content is requested, CCN router will search the name in a routing table. It is quite frequently occurred. This searching process may be boosted through parallel programing using GPU. Second, fast CCN routing table update method is required. The process of updating routing table fast is also important. This updating process may be boosted through parallel programming using GPU

How to solve Parallel Programming Design parallel algorithm to look-up Implement the parallel algorithm using CUDA GPU programming Optimize the program utilizing what we learned at EE817 lectures and surveys E.g. Aligned memory access, warp divergence, stream and so on Also, we will study new method to optimize the parallel process as possible We will use parallel programming method to solve these problems. We will design parallel algorithm to look up, implement the parallel algorithm using CUDA GPU programming, and optimize the program utilizing what we learned at EE817 lectures and surveys. For example, Aligned memory access, warp divergence, stream and so on. Also, we will study new method to optimize the parallel process as possible

Deliverable Implementation of conventional CCN routing table look-up Do not use GPU Implementation of GPU-based CCN routing table look-up Use parallel programming With large-scale table, we will compare the performance with some scenarios Here are our deliverables. Implementation of conventional CCN routing table look-up without using GPU. Implementation of GPU-based CCN routing table look-up with parallel programming. And we will compare the performance with some scenarios with large-scale table.

Task Allocation for Each Member Seyeon Kim Survey of CCN routing (method) Design parallel algorithm to look up routing table Junhee Lee Survey of CCN router (hardware) Implement conventional CCN routing method Both Investigate new methods to optimize the code Implement fast-table-look-up using parallel algorithm based on CUDA GPU programming Here is our task allocation. ~~ is for seyeon kim. ~~ is for me. We will do ~~ together.

Schedule May 18 : Proposal Presentation June 1 : Progress Report & Presentation June 22 : Final Report Submission 6/1 6/22 Here is schedule. weeks 1 2 3 4 5 6 Survey of CCN routing Survey of CCN router Design parallel algorithm Implement conventional CCN routing method Implement GPU-based CCN routing simply Optimize the code

References [1] Ericsson Mobility Report , June 2016 [2] Wang, Yi, et al. "Wire Speed Name Lookup: A GPU-based Approach." NSDI. 2013.