Magnet - A Dynamic Resource Management Architecture

MAGNET is an architecture  proposed in the PhD thesis by  Patty Kostkova  from   the former HiPeX research group in  City University , London, UK. Supervised by Dr. Julie A McCann.

The thesis is available in  postscript  and  compressed postscript . Also,  you can check our papers on MAGNET.

Overview

MAGNET is a new dynamic resource management architecture which meets the requirements of users in flexible and adaptive systems.

MAGNET enables dynamic trading of resources which can be requested indirectly by the type of service they offer, rather than directly by their name. A dedicated component, the Trader, matches requests for services against demands and establishes a component binding --- resource allocation. In addition, the architecture is extensible --- it does not constrain the information on services and allows user-customisation of the matching process. Consequently, this allows resource definitions to be parameterised (to include QoS-based characteristics), and the matching process to be user-customized (to perform QoS-based negotiation).
In order to fulfil the requirements of users relying on ever-changing conditions, MAGNET enables runtime adaptation (dynamic rebinding) to changes in the environment, constant monitoring of resources, and
scalability of the architecture.

Motivations

Computer systems no longer operate in centralized isolated static environments. Technological advances, such as smaller and faster hardware, and higher reliability of networks have resulted in the growth of mobility of computing and the need for run-time reconfigurability. The second factor significantly changing the role of a resource management is the inability of traditional operating systems to provide a flexible user-customisable platform where implementation of dynamic resource allocation strategies is feasible.

The dynamic management of this diversity of resources is the central issue addressed by the MAGNET architecture.

Applications in environments with frequently changing characteristics are required to participate in dynamic resource management, to adapt to ever-changing conditions, and to express their requirements in terms of quality of service.
 

MAGNET Architecture

MAGNET provides trading of service definitions based on a tuplespace paradigm. Here, we summarize how MAGNET actually supports the  requirements of applications in a mobile environment.
 

* Dynamic Trading.

  The Trader, based on a tuplespace paradigm, consists   of three components: the information pool, Trader operations on  tuples, and the matching operation. Dynamic trading is performed  by enabling components to define their services in terms of tuples  placed into the information pool (by operations Advert, and Bind),
  performing a matching operation and establishing a resultant binding.  Tuples can be withdrawn from the Trader by complementary operations,  (WithdrawC and WithdrawS).
 

* Extensibility.

 By enabling users to redefine tuple formats and user-customize the matching function, MAGNET supports extensibility.
 

* QoS-based Management.

  Extensibility and flexibility of the architecture enables QoS Management. Firstly, QoS Definition introduces QoS operators  (which can be user-customized). In addition, within the QoS Negotiation  phase, components can formulate their preferences by QoS-rating  operators and QoS-rating match in order to select the best tuple
  among a group of matching tuples. QoS Maintenance, based on QoS  Monitoring, enables two adaptation strategies to be implemented:  resource management and application adaptation.
 

* Dynamic Rebinding.

  Tuplespace-based implementation of four phases  of the rebinding process was described. In addition, we have discussed  all components involved in the process (Rebinders, Updater, Administrator)  as well as different situations into which the system may transforms  (as a result of different components performing the initiation and  renegotiation).

* Information Monitoring.

  Components for monitoring (Updater, Monitor) of service definitions placed into the information pool (for both parties --- clients and servers) ensure that component tuples are up-to-date at all times.

* Scalability.

  MAGNET supports operations Join and Leave to enable mobile users to use local resources transparently in a site where they have arrived. In addition, the architecture also supports scaling by enabling a tuple to be passed over the trading system to a particular Trader for processing using special components, Locators.

Publications