Browse Prior Art Database

Method And Apparatus For Deflating Resource Consumption In Object Oriented Based SOA Environments

IP.com Disclosure Number: IPCOM000200959D
Publication Date: 2010-Nov-01
Document File: 4 page(s) / 27K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a process which enables optimized resource usage during the service implementation phase in Service Oriented Architecture (SOA) based environments. The method decreases or prevents fragmentation which increases performance of the entire system

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 37% of the total text.

Page 01 of 4

Method And Apparatus For Deflating Resource Consumption In Object Oriented Based SOA Environments

In object oriented based Service Oriented Architecture (SOA) models, services are created and exposed to both external and internal consumers represented through service interfaces. Service interfaces can be defined through two approaches during the service specification phase (Note: SOA has service identification, specification, realization, implementation and monitoring phases):
• Top-down approach where the in and out parameters are defined initially in the interface and later the underlying objects get generated.

• Bottom-up approach where interfaces are generated by mix and match of underlying objects and later the interfaces are generated.

In both approaches, the generated service interfaces are exposed to consumers. The service interfaces may be exposed normally via two styles:
• In-line style where all the elements are strongly typed through the same interface; however, an option exists to refer to an element and its constellation by other elements.

• Offline style or standard style where elements are positioned in a library and individual service interfaces refer to the elements in the library.

During the service realization phase, developers import the service interfaces, refer to the library elements (in case of offline style)

which potentially generate all the objects

and create the implementation. Normally the byte memory stores such objects.

time the Object loader fetches this from the byte memory and creates instances in the heap based upon the input message. The instance may encompass a broad spectrum of highly varying elements to static elements. Treating this ecosystem uniformly imposes the following problems:
• Memory holes occur over a period of time and thus increasing the fragmentation in a heap (in the implementation phase of SOA) and resulting in poor performance as the defragmenter needs to switch across memory blocks which are far apart.

• RAM (both Stack and Heap required for the service) memory block allocated to the Services instances which are long running will eventually transfer to the memory blocks in the static memory (non -volatile memory) in a one to one fashion. This unnecessarily consumes the static memory resource because not the entire Stack memory is used at all times.

The disclosed method and apparatus addresses the above problems.

The performance of defragmenters is enhanced by logically partitioning the Heap into two or more parts and transferring only the actual used RAM memory (Stack) to the static memory.

An existing relevant solution is a method which focuses on organizing blocks as a file,

referencing the blocks through pointers, etc. It does not have differentiated treatment of

1

At run


Page 02 of 4

varying and static objects and different bands of Heap memory. Drawbacks of the current solution include:
• Disk usage is not optimal, thus poor performance
• Memory block tran...