Electrical computers and digital processing systems: multicomput – Distributed data processing
Reexamination Certificate
1998-12-24
2004-02-10
Le, N. (Department: 2858)
Electrical computers and digital processing systems: multicomput
Distributed data processing
C709S221000, C709S227000
Reexamination Certificate
active
06691148
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to distributed computing networks and, more particularly, to a system that provides quality of service across a distributed object-oriented computer network.
DESCRIPTION OF RELATED ART
The development and deployment of distributed applications has proliferated in recent years in part due to the creation of middleware. Middleware refers to a layer of software between the application and communication layers that offers a consistent, higher-level abstraction throughout the network. Middleware makes distributed applications easier to develop by facilitating communication between distributed applications.
An increasingly important category of distributed applications includes multimedia applications, such as video-on-demand. Multimedia applications demand high-perfornance communication layers. To meet these demands, communication layers offer features, such as quality of service (QoS) and multicasting, for multimedia applications to exploit. QoS refers to specific system performance requirements, such as the amount of a specific resource or group of resources devoted to satisfying a client (or caller) application's request. QoS allows for the reservation of guaranteed system properties. System properties include operational attributes, such as throughput and delay. Since these QoS features are offered only at the communication layer, new middleware must be developed to provide the application layer with access to these features.
Common Object Request Broker Architecture (CORBA) is middleware that was developed to facilitate such access. CORBA's interface description language (IDL) provides an important base for developing distributed applications because the IDL describes the programmatic interface to an object, including the type signature (i.e., a list of input and output parameters, along with the data types of each of the parameters) of the operations which the object embodies, independent of the underlying programming language, operating system, and communication medium. Through the programmatic interface, the IDL allows a programmer to develop a distributed application rapidly and transparently, without regard to the underlying services.
CORBA's IDL hides implementation details. As a result, distributed applications based on CORBA's IDL operate properly only as long as resources are plentiful. For example, distributed applications operate well where the programmers have prior knowledge of the system properties under which their code will execute and these properties remain relatively constant, such as where objects are either local (i.e., in the client's address space) or within the same local area network (LAN), and resources are plentiful.
In wide-area distributed environments, however, system properties are more dynamic, hostile, and less understood. This is a consequence of changing conditions, such as high delays, frequent failures, and low bandwidth, in the distributed system and the fact that programmers have no prior knowledge of these changing conditions.
As a result, usage patterns, QoS requirements, and underlying resources must be managed in order for a distributed application to operate properly over a wide-area network. Unfortunately, these features are precisely what is being hidden by the functional interface described by the IDL. To make a distributed application perform adequately, additional details regarding the design decisions embodied in an implementation of an object must be made available without sacrificing the software engineering gains, such as transparency, obtained by using object-oriented techniques.
A compounding factor that programmers face when developing distributed applications is that programmers are only used to handling the ideal system properties of local objects and not the dynamic system properties of distributed objects, such as their QoS properties. The system properties of local objects are considered ideal because the local objects do not fail independently of a client once they are created, and the delay for invoking a method is negligible.
Unfortunately, the system properties of distributed objects are far from ideal: they can fail unexpectedly and independently from a client, and the delay for a method invocation to return may be long due to communication overhead, for example. As a result, programmers typically include extra code with most invocations to remote objects to handle errors and performance conditions. Programmers find this coding very difficult to do, and the coding makes it even more difficult for programs to be used in an environment different from that for which they were originally hardcoded. Worse yet, the coding negates the benefits provided by the transparent interface definitions of the IDL.
Yet another factor that programmers encounter when they manage system properties in a distributed environment is that the information about the system properties becomes available at different times, in different locations, and from different sources. For example, information regarding system properties may come from many locations in a distributed system, such as from the client, the object, the communication layer, and the resources connecting them. The information may also come from many sources within the distributed system, such as the object designer, the client designer, the operations staff, and the end-user of the distributed application. If any of this information becomes unavailable for some reason, the distributed application will perform poorly.
Therefore, a need exists for a framework that supports QoS across wide-area and mobile environments to facilitate proper design and execution of distributed applications.
SUMMARY OF THE INVENTION
Systems and methods consistent with the present invention address this need through a quality objects (QuO) framework that includes a system of software components that provide a general purpose mechanism for the specification and satisfaction of application quality of service (QoS) requirements in a distributed, object-oriented computer software system. The QuO framework integrates knowledge of system properties over time, space, and source to facilitate proper operation of a distributed application.
In accordance with the purpose of the invention as embodied and broadly described herein, a system assures quality of service provided by a distributed network having at least one object. The system creates a contract that stores levels of quality of service offered by the network, determines a quality of service required by the object, and evaluates the contract to select a level of quality of service that corresponds to a current quality of service provided by the network. The system then compares the current quality of service to the required quality of service, and adjusts the current quality of service to obtain the required quality of service when the current quality of service fails to satisfy the required quality of service.
REFERENCES:
patent: 5713043 (1998-01-01), Baugher et al.
patent: 5898668 (1999-04-01), Shaffer
patent: 5944795 (1999-08-01), Civanlar
patent: 6003079 (1999-12-01), Friedrich et al.
patent: 6049819 (2000-04-01), Buckle et al.
patent: 6088732 (2000-07-01), Smith et al.
patent: 6154776 (2000-11-01), Martin
patent: 6226273 (2001-05-01), Busuioc et al.
patent: 6252857 (2001-06-01), Fendick et al.
patent: 6278693 (2001-08-01), Aldred et al.
patent: 6282581 (2001-08-01), Moore et al.
patent: 6330586 (2001-12-01), Yates et al.
patent: 6570867 (2003-05-01), Robinson et al.
David E. Bakken, On Specification, Metadata, and Binding of Multi-Property quality of Service, Proc. of 6thIntl. Working Conference on Dependable Computing for Critical Applications, IFIP, Grainau, Germany, Mar. 1997, 141-143.
David E. Bakken, Object-Oriented QoS: Some Research Issues, DARPA QoSA Meeting Presentation, 37thIETF Meeting, San Jose, CA, Dec. 1996.
David E. Bakken et al., QoS Issues for Wide-Area CORBA-Based Object Systems, Proc. of 2ndIntl. Workshop on Object-Oriented, Real-Time Depend
Bakken David E.
Loyall Joseph P.
Schantz Richard R.
Zinky John A.
Benson Walter
Le N.
Suchyta Leonard Charles
Verizon Corporate Services Group Inc.
Weixel James K.
LandOfFree
Framework for providing quality of service requirements in a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Framework for providing quality of service requirements in a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Framework for providing quality of service requirements in a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3301659