Research Achievements

1
Research Achievements

• Kenji Kaneda

2
Agenda

• Research background and goal
• Overview of my research achievements
• Phoenix
• Virtual Private Grid
• Summary and recent activities

3
Research Background and Goal
4
Background

• Grid computing
• Parallel computing with harnessing many
  widely-distributed resources
• E.g.) aggregate of PC clusters spread over
  multiple LANs

5
Traditional Parallel Computing vs. Grid Computing
Traditional parallel computing
Grid computing

• Reliable processors
• Single-LAN resources

• Unreliable processors
• Multi-LANs resources

6
Difficulty in Grid computing

• Frequent machine/network failures
• E.g.) 1 machine failure per a day
• Restricted Connectivity
• Administrative policies restrict communications
  between machines
• E.g.) firewall, NAT, DHCP

Firewall
TCP
Gateway
Gateway
7
Research Goal

• Allow a user to harness a computational grid like
  traditional parallel computing
• Fault tolerance
• Transparent communication on WANs

8
My Research Achievements

• Design/implementation of middlewares
• Phoenix
• Parallel programming library for accommodating
  dynamically joining/leaving resources
• Virtual Private Grid
• Command shell for utilizing hundreds of computers
  spread over multiple LANs

9
Phoenix
10
Phoenix

• Parallel programming library for accommodating
  dynamically joining/leaving resources
• Programming model for supporting migration of
  application states
• Transparent communication mechanism for WANs

11
Programming Model for Supporting Migration of
Application States

• Subsumes a regular message passing model
• Provides a namespace that does not depend on
  physical machines
• Programmer uses this name to specify a message
  destination
• Programmer can write a program without being
  aware of physical machines

12
Transparent Communication mechanism for WANs

• Overlay network construction
• Application-level message routing
• Processes can communicate with one another
• even if networks are not fully connected
• even if connection topologies change dynamically

13
Demonstration

• Boot processes on 3 subnets
• Add processes dynamically

14
Demonstration
15
Experiments (1/3)

• Speedup with fixed resources
• POV-Ray 78 speedup using 104 processors on 3
  LANs
• LU comparable to MPICH (on a single LAN)

16
Experiments (2/3)

• Speedup with dynamic resources
• POV-Ray takes advantage of dynamically added
  resources quickly

17
Experiments (3/3)

• Parallel shogi (Japanese chess) program
  on 720 laptop PCs
• 78 speedup

18
Related Work

• Grid enabled MPIs
• E.g.) MPICH-G G. Bosilca et al. SC02
• Based on a traditional message passing model
• Difficult to support dynamic changes of resources
• Communications libraries for Grids
• E.g.) Ibis A. Denis et al. HPDC04
• Static message routing

19
Summary Phoenix

• Parallel programming library for dynamically
  changing resources
• Good speedup with a large number of machines on
  multiple LANs

20
Virtual Private Grid
21
Virtual Private Grid (VPG)

• Command shell for utilizing hundreds of computers
  spread over multiple LANs

22
Features (1/2)

• User can submit jobs without caring
  administrative restrictions
• E.g.) cmd1_at_host1 cmd2_at_host2 gt file3_at_host3

Write to file3
host3
NAT
Firewall
Firewall
Execute cmd1
Execute cmd2
host1
host2
23
Features (2/2)

• Fault tolerance
• VPG can continue to run even if some machines are
  added/deleted dynamically
• No central server is required

24
Demonstration

• Environment
• 3 LANs
• CPU Sparc, x86, MIPS, PowerPC
• OS Solaris, Linux, IRIX

25
Demonstration
26
Related Work

• Grid job submission tools
• E.g.) Globus, Condor-G
• Difficult to submit jobs to machines under
  administrative restrictions

27
Summary Virtual Private Grid

• Command shell for utilizing hundreds of computers
  spread over multiple LANs
• Fast job submission to more than 100 machines

28
Summary and Recent activities
29
Summary My Research Achievements

• Middlewares for Grid computing
• Phoenix
• Virtual Private Grid

30
Recent Activities (1/2)

• Virtual SMP
• Emulates a multi-processor machines on a
  loosely-coupled computers

Virtual dual processor machine on two single
processor machines
31
Recent Activities (2/2)

• Virtual SMP
• Easy utilization of distributed resources with a
  common OS (e.g., Windows, Linux)