Adaptive bucket indexing mechanism to effectively manage...

Details Adaptive bucket indexing mechanism to effectively manage... Adaptive bucket indexing mechanism to effectively manage...

1999-11-24

2004-05-18

Homere, Jean R. (Department: 2177)

Data processing: database and file management or data structures

Database design

Data structure types

C379S201060, C379S012000, C707S793000

Reexamination Certificate

active

06738758

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to network management systems and, in particular, to provisioning services across a network or a set of networks.
2. Description of the Related Art
Network operators and network customers request the activation of services by use of simple service orders known as service activation requests (SARs), often provided on paper hardcopy. SARs are processed based upon customer demand and their requested activation date. Customers may demand immediate activation or delayed activation in the short- or long-term, i.e., service requests may have various levels of time criticality. In most situations this is a highly manual and labor- and paper-intensive process.
To add a further dimension to the problem, SARs (including service changes) may also involve complex interactions between many disparate network management systems (NMSs) and Element Management Systems (EMSs) and the elements therein and thus may not be confined to a single network resource under local control. Such a SAR is referred to as a complex SAR, containing many service sub-requests. For instance, a customer may request end-to-end service between two distant cities. Provision of this service may require obtaining dedicated bandwidth on several network resources owned and/or controlled by different parties (NMSs, or EMSs), such as a metropolitan area network in the originating city, a long-haul transport network, and a metropolitan network in the destination city.
Each SAR has its own “due date,” i.e., the calendar date and time on which the customer expects the entire service to be provided. All sub-requests share the same SAR due date.
This complexity introduces a level of difficulty in processing SARs with all of the necessary detail to meet customer requirements. As can be seen above, a large-scale (complex) service activation request necessarily includes numerous activation sub-requests addressed to network elements potentially owned by several different parties. Processing such a request currently over-burdens network operators, the network management system or systems where present, and the element management systems. Virtually all of current SARs and their components are handled manually using numerous communication systems, including computer networks and software systems, facsimiles, and the mails.
Prior art systems have attempted to address these shortfalls by a technique known as bulk provisioning. This technique requests multiple services in one large SAR, but is still essentially paper-oriented. While a single SAR contains numerous sub-requests affecting multiple network elements, the decomposition of the SAR is still performed manually, on paper, by human network planners. Bulk provisioning is thus insufficiently fast to keep up with the expanding needs of the modem network.
Automated bulk provisioning (ABP) is also known in the art. Typically, an ABP service activation module (SAM) contains many SARs for the same end-to-end service. For example, an ABP SAM might request “one thousand 25 Mbps virtual private network (VPN) services from San Jose, California to Austin, Texas.” Such a SAM thus consists of one thousand VPN SARs; as noted above, each SAR may contain many sub-requests. Current ABP systems, however, often lack robustness and tolerance for resource contention. In the present example, if any one of the sub-requests in any of the 1000 end-to-end VPN SARs cannot be processed, the corresponding SAR will fail. Any sub-request connections already provisioned for the failed SAR must be “rolled back” or de-allocated. The entire ABP SAM may also fail, if the management system that created it so dictates.
ABP systems are also non-scaleable to situations where a large number of different end-to-end services are requested because they lack robust means for managing resource overflow and provisioning errors.
Computer-based provisioning systems, including multi-threaded polling systems, have attempted to address the speed and efficiency problems, but with little success. Multi-threaded polling, in particular, requires too much memory and suffers from performance degradation when faced with complex SARs affecting multiple network elements. Resource shortfalls and the consequent contention issues also cause dead locks in which no SAR or sub-request is able to obtain the contended resource. Prior art automated systems also require complex rollback mechanisms to undo partially-filled complex SARs when a resource contention or error occurs.
What is needed is a method of automatically processing and decomposing complex SARs into service sub-requests appropriate to the required multitude of network resources affected. Furthermore, such a method must be fast, robust, and highly efficient and operable in a wide range of networks of varying complexity.
SUMMARY
Presently disclosed is a method for decomposing a service activation request (SAR) into a series of sub-requests, one per each specific network resource requested, and classifying the SAR with an adaptive bucket value (i.e., classifying it according to a ranking index that can change according to circumstances) based on the dynamic performance characteristics of the corresponding network resources. Each SAR has a virtual bucket value (ranking index) whose size is determined by the bucket values of its constituent sub-requests. Over time, as multiple SARs are processed, the bucket values associated with each sub-request's requested resource are adjusted to reflect system resource loading and availability. If the bucket value of a constituent sub-request is adaptively decreased down to a zero value, for instance due to an overflow condition or an excess of requests for that particular service, any new SAR requesting the same resource will be prevented from decomposition and processing. This will happen even if the sub-request is mature, i.e., due to be executed in the very near-term. This mechanism insures optimal utilization of network resources and provides an efficient, automated handling mechanism for processing complex service activation requests.
The present method also eliminates idling of the service activation request processing system due to resource contention and eliminates the need for complex database locking and/or blocking mechanisms to prevent over-utilization of network resources. Furthermore, the system operates without requiring any polling or absolute measurement of network performance in terms of the network's ability to provision services. Rather, the systems utilizes only the network response time to a request for services to determine its ability to handle further requests. By using a closed loop adaptation system, the present method constantly adjusts its response to each new complex SAR based on the ability of the system to process past SARs.


REFERENCES:
patent: 5195092 (1993-03-01), Wilson et al.
patent: 5528677 (1996-06-01), Butler et al.
patent: 5548761 (1996-08-01), Balasundaram et al.
patent: 5687224 (1997-11-01), Alley et al.
patent: 5835087 (1998-11-01), Herz et al.
patent: 5872779 (1999-02-01), Vaudreuil
patent: 5881131 (1999-03-01), Farris et al.
patent: 5883946 (1999-03-01), Beck et al.
patent: 5920846 (1999-07-01), Storch et al.
patent: 5953389 (1999-09-01), Pruett et al.
patent: 6005926 (1999-12-01), Mashinsky
patent: 6029195 (2000-02-01), Herz
patent: 6167050 (2000-12-01), Chung
patent: 6226365 (2001-05-01), Mashinsky
patent: 6249578 (2001-06-01), Gilles et al.
patent: 6314093 (2001-11-01), Mann et al.

Affiliated with

Also associated with

Rating

Adaptive bucket indexing mechanism to effectively manage... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Adaptive bucket indexing mechanism to effectively manage..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive bucket indexing mechanism to effectively manage... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3269619