Fabio Kon Tomonori Yamane Christopher Hess

1
Dynamic Resource Management and Automatic
Configuration of Distributed Component Systems

• Fabio Kon Tomonori Yamane Christopher
  Hess
• Roy Campbell M. Dennis Mickunas
• Departments of Computer Science
• University of São Paulo
• University of Illinois at Urbana-Champaign
• http//choices.cs.uiuc.edu/2K

2
Introduction

• Modern Society requires software developers to
• produce large quantities of programs
• write large, complex programs
• support different OSes
• support different machine architectures
• Partial solution Component Technologies
• Enterprise Java Beans
• ActiveX Controls, COM
• CORBA Component Model

3
Problems in Existing Component Technologies

• Lack support for representing dependencies among
  components
• Difficult to support
• Automatic Configuration
• Dynamic Reconfiguration
• Fault-tolerance
• Adaptation, etc...

4
Lack of Proper Dependence Management in Existing
Operating Systems

• 1. Administration / Configuration
• Junk libraries left on Windows after uninstall
• 2. System Architecture
• Different (static) instances of same OS
• Configuration of Microkernels
• 3. Fault-tolerance
• Module failure not handled by others

5
Our Solution

• Infrastructure for Dependence Management
  supporting
• Automatic Configuration
• Dynamic Reconfiguration
• Code Distribution
• Dynamic Resource Management
• Help developers to support
• Fault-Tolerance
• Consistent Reconfiguration
• Adaptation

6
Automatic Configuration Service

• Automatically instantiates applications and
  services by assembling their components.
• Based on
• Prerequisites static representation of
  dependencies.
• ComponentConfigurators dynamic representation
  of dependencies.

7
Prerequisites

• What a component needs to run
• nature of hardware resources
• share of the hardware resources
• software services (i.e., components) it requires
• Video Client example
• PC with Sound card
• 50 of CPU gt400MHz
• MPEG decoder, CORBA Video Service

8
Simple Prerequisite Description Format (SPDF)

• Video Client application
• hardware requirements
• machine_type x86
• os_name Linux
• os_version 2.2
• min_ram 5MB
• optimal_ram 40MB
• cpu_speed gt400MHz
• cpu_share 50
• software requirements
• Decoder CR/video/decoders/MPEG
• VideoServer NS/local/MPEGVideoServer

9
Automatic Configuration Process

• 1. Fetches component code and prerequisites from
  the Component Repository.
• 2. Dynamically links component code into the
  application address-space.
• 3. Based on the prerequisites, repeats the
  process for other components.

10
Automatic Configuration Architectural Framework
fetch prerequisites
load application
Component Repository
Prerequisite Resolver
fetch components
return reference
Prerequisite Parser
QoS-Aware Resource Manager
Cache
11
Component Configurators

• Reify dynamic inter-component dependencies.
• Created on-the-fly by the Prerequisite Resolver.
• System and application software can inspect and
  reconfigure the Dependence Graph.

12
ComponentConfiguratorFramework

• Allows browsing, inspection, and reconfiguration
• Can be customized through inheritance
• Clear separation of concerns

13
QoS-Aware Distributed Resource Management

• Global Resource Manager (GRM)
• one in each cluster
• maintains an approximate view of the cluster
  resource utilization
• Local Resource Manager (LRM)
• runs in each node
• exports the state of the local resources
• Has a Real-Time Scheduler (DSRT)
• admission control, reservation, and scheduling

14
Loading an Application with the Resource
Management Service

• 1. Client contacts local LRM, giving application
  name and QoS requirements
• 2. LRM performs admission test
• 3. Request forwarded to GRM
• 4. GRM forwards request to best candidate
• 5. Remote LRM performs admission test,
  reservation, and runs AutoConfig.

15
Experimental Results

• Testbed
• 2 Sun Sparc Ultra-60, two 360MHz CPUs
• Solaris 7 OS
• 100Mbps Fast Ethernet

16
AutoConfig ServiceLoading Several Components
17
AutoConfig ServiceLoading Components of
Different Sizes
18
Related Work

• Automatic Configuration
• Customizable Operating Systems
• Java/Jini
• Prerequisites
• SOS operating system Shapiro 94
• QoS description languages Frølund 99
• Distributed Resource Management
• Globus Foster and Kesselman 98
• Legion Grimshaw 97

19
Future Work

• Apply the framework to the Gaia project (Active
  Spaces, Ubiquitous Computing).
• Enhance version control
• Example component A v. 1.2 may require
  (component B v. 2.x) or
  (component B v. 1.x and component C v. 2.3)
• Performance Analysis of the Resource
  Management Service.

20
Conclusions

• As computing devices become pervasive in our
  society, we will encounter
• highly dynamic, heterogeneous environments
• complex dependencies
• difficult management
• 2K presents an integrated architecture that helps
  managing this complexity in a clean and efficient
  way.

21
How to Contact Us

• Fabio Kon
• kon_at_ime.usp.br
• http//www.ime.usp.br/kon
• Christopher Hess
• ckhess_at_cs.uiuc.edu
• http//choices.cs.uiuc.edu/2k

22
(No Transcript)
23
Mobile Configuration Agents

• Suitable for Large-Scale Systems
• Useful for
• code distribution
• dynamic reconfiguration
• inspection

24
The 2K Architecture
25
Overall Architecture(relationships)