 |
Proactive
Application Management System (PAMS)
Principle Investigator:
Graduate Students:
Objective
Management of large-scale Network-Centric Systems (NCS) and their applications
is an extremely complex and challenging task due to factors such as centralized
management architectures, lack of coordination and compatibility among
heterogeneous network management systems, and the dynamic characteristics
of networks and application requirements. The
goal of this research is to develop an integrated framework to achieve
end-to-end intelligent and proactive management system that can be used
to manage large-scale network centric systems and their applications. This
framework will provide the ability to write management programs to manage
any required function or property (performance, high assurance, fault,
quality of service, etc.) of the network-centric systems and their applications
during all the phases of their operations. Our ultimate objective
is to consider the management of network-centric systems and applications
starting from the design phase and forward rather than being after thought
process. In this thesis, we present a framework to identify the management
functionality and develop proactive and adaptive management services and
an implementation of a Proactive Application Management System (PAMS) based
on that framework.Our
implementation approach utilizes delegated mobile agents to implement the
management functions required by any network-centric system and/or application.
We also present experimental results and evaluation of the management services
offered by the PAMS prototype. The experimental results demonstrate
that our agent-based approach can lead to significant gains
in the performance and low overhead fault management of parallel/distributed
environment.
Research Description
End-to-End Integrated Management Framework
The management framework we are developing can be viewed in terms of three systems: Network and Protocol Management (NPM), Management Computing System (MCS), and Application-Centric Management (ACM). The NPM is responsible to collect management information not only about the network devices, but also information related to computer processes, file systems, user access information and patterns, and protocols. The NPM will also perform tasks to manage the network devices, protocol functions, computer processes and file systems. The MCS provides the core management functions to manage the whole system resources from system perspective rather than component level perspective. In order to achieve that, the management information collected at the lower level (NPM) will be analyzed and abstracted into suitable data structures or format to perform efficient system level management functions. The MCS design concept is analogous to the operating system in computing systems. The operating system manages the computing system resources (memory, I/O, CPU, and processes). Similarly, the MCS acts as an automatic system manager that provides management functions to achieve application centric management tasks.
The ACM provides two main functions: Assist
in the development of application management routines, and provide intelligent
proactive management for a wide range of network applications. Figure 1
shows a block diagram of the proactive end-to-end management framework.
In what follows, we describe the main components of each layer in this
framework.
Proactive Application Management System (PAMS)
The ACM Service provides
the user with the tools required to describe and characterize the management
requirements of any network-centric system or its application. The MCS
provides the management services to automatically configure the application
or system resources, monitor and control the execution of an NCS application.
The NPM service provides the appropriate interface to existing network
management systems and utilize their services in order to proactively manage
and control the operations of the NCS or its applications.
The
PAMS has been implemented within our JINI/Java framework for integrated
analysis, control, and management of network applications as shown in Figure
2. The Application Management Editor (AME) provides application
developers with the services required to write management routines to maintain
application quality of service requirements, performance, fault management
strategy, or security management policy. Once the application management
routine of an application is described using AME, the next step is to utilize
the management services provided by the Management Computing System (MCS)
to build the appropriate Application Execution Environment (AEE) that can
dynamically manage and control the allocated resources to maintain the
application requirements. The MCS utilizes a JINI Server that provides
a scalable and open architecture to dynamically and seamlessly add and
remove network management services as well as application management services.
The JINI server provides two main services: Discovery Service to register
and remove network management services, and Lookup Service to determine
the appropriate site to obtain any requested network/application management
service. Once all the management requirements are identified, the MCS assigns
one Application Delegated Manager (ADM) (see Figure2)
to manage one or more application attributes (performance, fault, security,
etc.). For each task in the application, the ADM launches an appropriate
Task Agent (TA) that monitors the task execution and stores the task states
in a Task Information Base (TIB). The TA monitors the task execution using
appropriate data structures that we refer to as Sensors (S) and intervenes
whenever the task execution on the assigned resource cannot meet its requirements
using the task Actuators (A). The task actuator can suspend, save task
execution state, or migrate the task execution to another remote machine.In
what follows, we describe our methodology to proactively control and manage
any application or resource attribute or functionality (performance, fault,
security, deployment, QoS, etc.).
Figure 2
The Management Architecture of PAMS
Publication
1.Yoonhee Kim, S. Hariri, M. Djunaedi, ?Evaluation of PAMS Adaptive Management Service?, HCW 2000 Ninth Heterogeneous Computing Workshop, Cancun, Mexico, May, 2000
2.S. Hariri, Yoonhee Kim, ?Design and Analysis of a Proactive Application Management System (PAMS)?, IEEE/IFIP 2000 Network Operations and management Symposium (NOMS2000), Hawaii, April, 2000
3.Yoonhee Kim, S. Hariri, M. Djunaedi, ?Experimental Results and Evaluation of the Proactive Application Management System (PAMS)?, 19th IEEE International Performance, Computing and Communications Conference (IPCCC 2000), Phoenix, February, 2000
4.Yoonhee Kim, M. Djunaedi, S. Hariri, ?Toward Proactive Application-centric Management?, 1999 International Conference Parallel and Distributed Processing Techniques and Applications (PDPTA '99), Las Vegas, June 1999.
5.Yoonhee Kim, S. Hariri, ?ExNet: An intelligent network management system using Internet technologies?, WebNet98--World Conference of the WWW, Internet and Intranet, Orlando, November 1998
6.S. Hariri, Yoonhee Kim, P. Varshney, R. Kamiski. D. Haugue, C. Maciag, ?The End-to-End Proactive Network Management?, IEEE/IFIP 1998 Network Operations and management Symposium (NOMS98), New Orleans, February 1998
Demo
PAMS
Demo