Configuration engine

Description of equivalent WO 2004021109 A2

CONFIGURATION ENGINE

Inventor: Jeffrey F. Bryant COPYRIGHT NOTICE

[0001] A portion of the disclosure of this patent document contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

RELATED APPLICATIONS

[0002] This application claims the benefit of U.S. Provisional Applications No. 60/407,555, filed August 29, 2002, and No. 60/406,853, filed August 29, 2002. This application also claims the benefit of U.S. Applications No. 10/290,939, filed November 8, 2002 and 10/303,268, filed November 25, 2002. Each of these applications is herein incorporated in its entirety by reference.

BACKGROUND OF THE INVENTION

TECHNICAL FIELD OF THE INVENTION

[0003] The present invention relates to software architecture. More particularly to software architecture incorporating configuration engines for the management of multiple processing nodes.

BACKGROUND

[0004] The rapid evolution of technology has posed significant problems, as well as benefits. Some technologies never achieve their full potential while others evolve rapidly, leaving earlier versions obsolete shortly after they have been installed. Technologies may need to be frequently substituted or otherwise adapted to compensate for different needs. Software particularly must be made amenable to

l - substitution and adaptation and can be a means of allowing integration of new hardware or allowing existing hardware to fulfill new functions.

[0005] Large-scale software development has evolved rapidly from its inception. Through the 1980s large-scale software was developed in modular systems of subsystems. Even today these are the most common systems in use. These systems are largely hardware dependent, problems or errors could be detected down to the level of the subsystem. These systems were based on point solutions where the problem/solution is functionally decomposed into subsystems. Potential reuse of the software for other applications must be anticipated during development and integrated into the software design. Extensions of the software are difficult and can only be achieved when such extensions were anticipated and designed into the system architecture itself.

[0006] In the 1990s, some improvement came with the advent of Object Oriented Systems (OOS). Object Oriented Systems were still deficient in a number of respects. OOS are still hardware dependent, they are designed for specific hardware configurations and modules are not productized. Off the shelf components could not be easily integrated in to a software system since each piece of software was developed for a particular hardware platform using different languages. No standard interface was available. These systems were based, like their predecessors, on point solutions. The point solutions for OOS are derived using Object Oriented Analysis. Extension of the system using existing components was difficult as a result of the multiplicity of languages used.

[0007] In recent years, research and development has centered on layered or component based systems. In such a system a thin common layer or component base class is used in the development of all software modules. Each of the major capabilities of the system is represented by at least one module or component. These modules or components are thus "wrapped" in the thin common layer. Independent components are developed, tested, and packaged independently of each other, and while operating have no knowledge of their environment, since all input/ output is constrained to interface ports connected from the outside. Run time discoverable parameter ports control specific behavior.

[0008] Component technology has in recent years become an area of increasing interest given the above challenges. Component technologies such as, CORBA, Common Object Request Broker Architecture (developed in 1997), allow for increased flexibility when implementing business processes. By combining components many different software products can be created from existing modules. This increases the speed and efficiency of software development thereby better meeting client and internal demands and decreasing costs associated with development.

[0009] The goal now is to make software components that allow reuse by performing a particular function and providing an appropriate interface with a larger system. Each component should be autonomous regarding its particular functionality. This autonomy would allow changes to be made with individual components without disturbing the configuration of the entire system. Relating the various quasi- autonomous components to each other results in a high degree of complexity in communication and synchronization code.

[0010] A system of reusable and flexible components would be especially useful for developers of large and complex software packages, such as military contractors. In the past, software was designed specifically for a contract. When a new contract was bid for, the contractor started from scratch. As discussed above, differences in language and architecture prevented different functionalities from being reused from earlier contracts. Since the software was newly developed there remained a relatively high risk of failure in the software or its interfaces, therefore the new software required testing and packaging, adding to the cost of the contract. The application of a flexible framework of reusable and interchangeable components would enable a client to leverage earlier development investments and minimize risk of failure in the development process. Contractors would be able to provide clients with more accurate and lower bids and possibly prototypes or catalogues of products easily configured to the clients needs.

[0011] A similar, though different, architecture is SCA- Software communication architecture. This architecture is used in such applications as SDR (Software Defined Radio). SCA has specific IDL interfaces defined for software radios. Any new desired capabilities must fit in to pre defined IDL. SCA provides an interface framework; as such, it is not hardware independent. While peer- upper layer interfaces are well defined in SCA, lower layer interfaces are largely ignored. Another disadvantage of SCA for more general application is its total reliance on CORBA layered communications. Such problems present themselves in CPU overhead and quality of service. Messages can be delivered out of order and processed by different threads when belonging to the same data streams. Thus the SCA architecture is unsuitable for the distributed computing application.

[0012] Rocray et al. in a published US Application Pub. No. US 2002/0065958 Al disclose a multiprocessor system that comprises a plurality of processor modules, including a software management processor, a non- volatile storage memory configuration (NVS), and a plurality of software components stored on the NVS configured for use in the processor modules. The application further discloses a software generic control information file used by the software management processor to relate the compatibility of software and to determine which of the software components to distribute to a processor module that requires software stored on the NVS.

[0013] In published PCT application, WO 02/057886 A2, Talk2 Technologies discloses methods, systems, and computer program products for dynamically accessing software components in an environment that may include a plurality of processing nodes. In the '886 reference, each node includes one or more software objects, such as one or more software component objects (virtual processors), a controller object, a database object, a trace object, an agent object, etc. Requests for the functionality implemented by a particular software component are load balanced across the available instances. If no software components are available, a request may be submitted to a scheduler. A software component also may be reserved for future processing. Relationships between software components are defined by platform independent logic that is accessible through a database object and processed by a controller object. An agent object tracks which software components are available at the one or more nodes for which the agent is responsible.

[0014] In order for such a component system to properly function a central infrastructure must provide a forum for this communication and synchronization for components and control the allocation of the tasks to the various components based on the capabilities and availabilities of those components, thereby preventing conflicts and redundancies.

[0015] Clearly what is needed to create a flexible framework of reusable and interchangeable components, developed and working independently, coordinated by controls that can be manipulated without substantial reengineering or programming in runtime.

SUMMARY OF THE INVENTION

[0016] The invention is devised in the light of the problems of the prior art described herein. Accordingly it is a general object of the present invention to provide a novel and useful technique that can solve the problems described herein.

[0017] One embodiment of the present invention provides a system for management of a distributed computing system including a tasking manager, whereby a plan is formulated describing a task to be carried out by the distributed computing system and detailing resources required for execution of the task a framework manager, communicating with the tasking manager and comprising at least one configuration manager, including a library within which are disposed at least one configuration template a means for ascertaining the current configuration of the distributed computing system a component assignment unit communicating with said tasking manager, the library, and the means for ascertaining the current configuration; and an output whereby the configuration manager directs the allocation of tasks to available system resources.

[0018] In such a system for management of a distributed computing system the configuration template may detail a tested configuration compliant with the distributed computing system and the component assignment unit may be configured to: compare said required resources with said each configuration template in said library to select usable configuration templates; compare said usable configuration templates with said means for ascertaining the current configuration; determine if system resources required by said usable templates are available; and select, based on comparisons of said usable templates with said current configuration, an optimal resource allocation for execution of said task.

[0019] According to one embodiment, system resources may include software processing units and hardware components. Such system resources may be adapted to RF applications such as tuners, channelizers, transmission security components, delay/multi-band radios, modems, communication security components, and decoders.

[0020] A further embodiment of the present invention provides a configuration manager for the assignment of task elements to processing units in a component based distributed system, including a library wherein are disposed at least one configuration template; a means for ascertaining currently available system resources of the distributed computing system; a component assignment unit communicating with a tasking manager, the library, and the means for ascertaining currently available system resources; and means for communicating assignments of the task elements to the processing units.

[0021] In this embodiment, the component assignment unit may be configured to: compare said required resources with said each configuration template in said library to select usable configuration templates; compare the usable configuration templates with the means for ascertaining the current configuration; determine if system resources required by the usable templates are available; and select, based on comparisons of the usable templates with the current configuration, an optimal resource allocation for execution of the task.

[0022] According to one embodiment, system resources may include software processing units or hardware components. Such system resources may be adapted to RF applications, such as tuners, channelizers, transmission security components, delay/multi-band radios, modems, communication security components, and decoders.

[0023] Another embodiment of the present invention provides a method for the configuration of components in a component based distributed system including the steps of: obtaining a plan schematic; obtaining at least one configuration template from a configuration template library; matching the configuration template with system requirements defined by the plan schematic; generating a preliminary result from the step of comparing the system requirements with the configuration template; obtaining a current configuration of components in the component based distributed system describing currently available system resources; matching the current configuration with the preliminary result; generating a final result of the step of matching the current configuration and the preliminary result; and communicating the result to distributed processing centers. [0024] According to one embodiment, configuration templates include tested configurations for the component based distributed system. The preliminary result may include a plurality of useable templates while the final result is an optimal configuration for the components given said current configuration of the system.

[0025] According to one embodiment of the present invention, the method is executed within a framework manager. A tasking engine may formulate the plan.

[0026] According to one embodiment of the invention, matching the current configuration with the preliminary result is undertaken so as to obtain a template with an optimal correlation between the requirements of the template and resources available to the system.

[0027] According to one embodiment of the present invention, components may include software processing units or hardware components. Such components may be adapted to RF applications

[0028] Still other objects and advantages of the present invention will become readily apparent to those skilled in this art from the following detailed description, wherein we have shown and described only a preferred embodiment of the invention, simply by way of illustration of the best mode contemplated by us on carrying out our invention. As will be realized, the invention is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The present invention will be readily understood by the following detailed description in conjunction with the accompanying drawings and exhibits, wherein like reference numerals designate like structural elements, and in which:

[0030] Figure 1 is a flow chart of one embodiment of the present invention.

[0031] Figure 2 is a flow chart of the tasking process utilizing one embodiment of the present invention.

[0032] Figure 3 is a flow chart of a sample plan constructed by a tasking engine within a framework implementing one embodiment of the present invention.

[0033] Figure 4 is a flow chart of one embodiment of the present invention and its relationship to components and client processors.

[0034] Figure 5 is a block diagram of the configuration manager according to one embodiment of the present invention.

[0035] Figure 6 is a follow chart illustrating the steps in the configuration method according to one embodiment of the present invention.

DETAILED DISCRETION OF THE PREFERRED EMBODIMENT

[0036] The deployment of components to solve a processing problem may be performed in two steps. The first describes what the elements are and the logical connection between them. The second step involves placing the components on processing elements such that the performance of all the components is guaranteed. This two- step technique allocates components to processing elements using a set of known working configurations that are described in a database. The algorithm queries each processing element for the template database and a set of rules that optimize the processor utilization and minimize the disruption to existing components.

[0037] One embodiment of the present invention is a component manager or engine integrated into a SIMPLE, script based, task engine. A user can input a plan of what components are necessary for a particular procedure via an API. Parameters for this can be set and controlled without changes being made to code. Such a script-based system requires little effort from the user to coordinate even complex tasks.

[0038] Referring to figure 1, this figure illustrates the operation of a configuration manager operating according to one embodiment of the present invention. The configuration manager or engine uses configuration templates to describe the possible combinations of processors capable of performing a particular function. These templates are stored in a database and are recalled by the configuration engine. The configuration engine maintains a current configuration of the system. The current configuration consists of a list of: the processors, of components running on each processor, current memory and CPU utilization of the processor, current interconnects between processors, and current loading of the processor.

[0039] This configuration is compiled dynamically at run time just before a new configuration is made. The configuration manager or engine is then able to compare the configuration template, the current configuration, and requirements of the task or schematic, inputted by the user and comprising: the list of components, the connections between each of the components and the component parameter values and other configuration information.

[0040] The configuration manager or engine then assigns the task to the processor best suited to its requirements and without conflicts and maintains a database of what components are loaded on which processors an monitors the processor loading (Memory, CPU and Input/Output bandwidth utilizations).

[0041] The configuration manager when assigning tasks to processors applies a component assignment algorithm designed to efficiently deploy components on as few processors as possible. Secondarily, the algorithm seeks to avoid disruption to currently executing plans or tasks. The algorithm avoids loading components until the optimal allocation is obtained. If the loading process fails, no assignments are actually made.

[0042] One embodiment of the present invention is best understood in the context of a Signal Intelligence/Information Warfare (SIGINT/IW) System, a component is an object-oriented set of reusable and new software, and if required compatible hardware, that performs a specific system task.

[0043] The present invention allows for the development and integration of custom object-oriented components that satisfy the assigned task(s) and are scaleable, modular, extensible, platform independent, open, robust, and efficient.

[0044] The general approach in this software development is to use object-oriented technology to model and build software, using the Unified Modeling Language (UML). This development technology utilizes the UML semantics and graphical notation to express the software design. [0045] In one embodiment, the Rational Rose application is the primary UML tool used to specify, visualize, construct and document the elements and artifacts of a component consistent with the preferred embodiment of the present invention. Rational Rose is used to implement UML to create a blueprint or model for object- oriented system and software application construction.

[0046] An embodiment follows a set of standards to generate the object oriented design and software for the framework and components.

[0047] Once the component software model is reviewed, updated and stabilized the detailed lower level classes and internal sequences are coded to complete the component Visual C++ source code. The component source code is then compiled and tested.

[0048] Through the use of Object oriented technology, the component developer creates the application by extending the component base classes. Most methods of these classes have desired default behaviors, and need not be extended for simple applications. Component development can proceed in an iterative approach, with each iteration extending more and more of the classes and overwriting more and more of the base class methods, increasing the component maturity until the ultimate desired behavior and performance is attained.

[0049] A component, according to the present invention, interfaces and exists with a framework having been optimized for throughput, and will continue to have advances in performance with successive releases. The small penalty in run-time performance is greatly offset by the reduction in non-recurring engineering.

ARCHITECTURE OVERVIEW

[0050] One embodiment of the present invention is an Information Dominance Systems product family for new SIGINT/IW Systems. The embodiment has a component- framework architecture that consists of a framework of expandable sets of reusable software components with compatible COTS hardware. A framework makes components work together, provides a common run-time infrastructure for components, and provides a common set of base classes for component developers. One embodiment is an object-oriented system and software development approach for an evolving scaleable family of products. One embodiment is Operating System and platform independent in that the architecture is designed to run on PC WindowsNT, Sun Solaris, and embedded platforms such as VxWorks and MC/OS.

[0051] The preferred system provides a Product Store service, a repository for SIGINT products, accessible as either objects or as database entries. The system developer may extend the Product Store as required.

[0052] The component-framework architecture of the preferred embodiment provides common standard interfaces to the components. This facilitates testing at the component level. An embodiment of the present invention provides programmable component test capability. This capability consists of TestSource, a programmable signal and event source component, TestSink, a controllable signal and event logging and display component, and the MAUI (Maintenance And User Interface), which provides a graphical user interface into the component interfaces of individual components, processors, plans, and systems.

[0053] The architecture of the preferred embodiment of the present invention is component framework architecture. Components "Do the work" of the application. A component encapsulates the signal and data processing inside common (extensible) interface. Frameworks make components work together. The framework provides a common upgradeable run-time infrastructure for components and provides a common set of base classes for component developers. FRAMEWORK MANAGER

[0054] Based on the mission, priority, and component requirements a dynamic plan is constructed, by a tasking engine, to accomplish the individual component task(s). The tasking engine issues instructions to a framework manager. The framework manager manages the deployment of components to remote processors through a processor manager. The framework manager contains, in the preferred embodiment, a configuration manger. Some components may be capable of running some tasks, incapable of running others and capable of running still others yet with sub-optimal results. Components may already be assigned tasks, and have insufficient capacity remaining to execute additional tasks. The configuration manager seeks to rectify this. The configuration engine gathers information stored in a configuration template associated with each of the components, interrogates the processors about their current status and compares the results to the dynamic plan developed by a tasking engine. The configuration engine then assigns tasks to processor managers based on the results obtained. Thus an optimum allocation of resources is achieved even before a task is assigned to a processor or component. The framework manager then no longer is in direct control of the task, leaving the management thereof to the processor manager.

PROCESSOR MANAGER

[0055] The processor manager is application independent. A processor manager is associated with one or more specific components. It takes commands from the framework manager and manages the operation of components local to the processor.

COMPONENT(S)

[0056] Each component manages its local transforms and its associated I/Os. Component interaction is fully decoupled from other components. Any component interaction is managed by the Framework Manager and an associated Processor Manager. Some components such as Channelize, Detect, Copy, and Recognize are application independent. [0057] The framework manages the deployment of components to processors in the system. The framework handles the management of interconnection of components. The deployment and interconnection of components may be done individually, or through the submission of a plan, containing this information.

[0058] Task applications provide translation of system operator or user required services to plans, component deployments, and component parameter settings. This includes copy (providing a software radio), search (a SIGINT signal acquisition and recognition system), and others. These task applications are extensible by system developers.

COMPONENTS

[0059] Components are essentially small subsystems, with common interface characteristics and expected (control) behavior. By using Object-Oriented technology, these common interfaces and behaviors can be provided to the developer and enforced through inheritance. The component encapsulates the signal and data processing inside a common interface. The framework provides deployment, communications, and health monitoring services to the components.

[0060] The interfaces and expected (control) behaviors of components constant with the present invention are provided for the developer. The developer can focus his/her effort on the signal and data processing requirements of the component, and benefit by reusing the deployment, communications, and engineering services of the component base classes.

[0061] The component interface is designed not to be extended. The component base classes are extended "inside" the component. "Specific Interface" provides an interface extension mechanism for special cases. [0062] A component implements a single function or small sub-system capability, for example, detection, base-band tuning, signal recognition, etc.

[0063] A component accepts input and produces output, generally in the form of signals (e.g. stream of complex floating point samples) or events (detection "alarms").

[0064] A component developer is generally unaware of the existence or behavior of other components. This frees the developer to solve the signal or data processing problem assigned, minimizing the development of interfaces and system level interactions. Interfaces and system level interactions consume most of the software and system development activity, while providing no perceived value to the customer.

[0065] A component is controlled using run-time discoverable parameter sets. These parameter sets are associated with both the component controller and the output ports. Applications external to the component may obtain the current parameter values and definitions, and may change the parameters of the component. The component is notified of changes and may behave as the developer programs it to.

[0066] A component is developed independent of the target processor platform. The component developer need not understand the details of the operating system or machine architecture to create functioning, high performance components. The details of the operating system and target platform are insulated, as the developer uses an Operating System API for the services required. The developer typically uses a user-friendly development environment such as MicroSoft Visual C++ for initial code and debugging, and then rebuilds the component for the target.

COMPONENT BASE CLASS SERVICES

[0067] The Component Base Class provides the required services and interfaces for the component developer. These include: Component

Provides all component services (through delegation) Creation, destruction, stop, start, shutdown, provide ports, plots, identity Control

Register for sending, receiving Component Interface

Access to services to the specific component

Get output port and port info, set/get parameters, parameter change service, enable message logging, identity, statistics registry, send, get specific interface InputPort

Input for signal and event data Register, accept in (by type), identity, status, get input OutputPort

Output for signal or event data

Send, identity, parameters of processing for this output, set/get type Parameters

Self-Describing interface for control for processing, output ports, plots Also used for other self-describing interfaces Set, get, get definitions, verify, register for update OSAPI

Operating System API Time, time-related services, spawn, identity Plotlnterface Graphical debug

Add, remove, get, set, get definition, set parameters, plotfx, plotxy, text, clear Transform

Base class of one instance of signal processing Accept signal or event, parameters, output control Message Logging

Severity managed error logging logMessage

SETTING UP THE DEVELOPMENT ENVIRONMENT

[0068] In an embodiment setting up the development environment is pre-requisite to component development and consists of the configuration of individual WindowsNT workstations connected to the developer's company LAN.

EXAMPLE OF WINDOWS NT WORKSTATION BASE CONFIGURATION

[0069] Microsoft Windows NT 4.0 with Service Pack 4 (or later)

[0070] Intel Pentium III 600 MHz Processor (512 cache)

[0071] Minimum configuration must include: i. Intel Se440BX2 motherboard with Ultra ATA IDE, 440BX2 chip set, ii. PCI bus, 3 dedicated PCI slots, 1 dedicated ISA slot, 1 shared iii. ISA/PCI slot, AGP graphics slot iv. 512M bytes ECC Memory (DIMM) v. Western Digital 20GB IDE hard drive AC28400 vi. 1.44Mb 3 ^lA floppy drive vii. enhanced IDE CD drive viii. Intel Pro/100 Plus Ethernet card ix. Matrox G400 w/ 16MB video card x. Monitor 21" (Minimum recommended resolution 1600x 1200) xi. 320 Sound Card xii. PS2 104 key Keyboard xiii. Microsoft PS2 IntelliMouse [0072] Fig. 1 is a block diagram of elements of a framework manager 100 configured according to one embodiment of the present invention. A tasking engine provides a new plan or schematicl02. The plan details the components and their connections, but does not assign processors. The configuration manager or engine 104 is in charge_^ of assigning components to processors. To do this the configuration manager 104 first inspects the plan or schematic 102 looking for unallocated components. If all components are allocated, it then checks to see is each component is capable of running on the processor for which they have been configured.

[0073] The configuration manager 104 queries each process manager looking for components that have already been loaded. It constructs a database 106 for each processor containing a list of those components already loaded. This database 106 contains the current configuration of the system.

[0074] The configuration template then loads the template database or library 108. The template library is compared with the plan, and those templates that provide viable configurations are collected. These viable templates are then compared with the current configuration. The configuration manger 104 is configured to minimize the resources dedicated to a particular component. Thus the optimal assignment is one that minimizes the number of processors devoted to a particular component.

[0075] The configuration manager 104 assigns components first to empty processors, then to partially full processors. The configuration manager then updates the template library 108, including the new configuration as a template.

[0076] If the component is successfully loaded onto its assigned processor, the processor manager sends a success message back to the configuration manger 104. If it fails to load, the configuration manger is alerted to the error.

[0077] Referring now to Fig. 2, which is a block diagram of the tasking/ configuration process constructed according to one embodiment of the present invention. The plan 102, formulated by a tasking engine 110 is deployed on a fabric of processor units 112. This is realized through the framework manager 114 acting through the processor manager(s) 116. The processors then execute the plan 102, thereby creating the product 118.

[0078] Fig. 3 illustrates one possible plan characteristic of an RF (radio frequency) application configured according to one embodiment of the present invention. A Wide band tuner 202 is coupled with a channelizer 204 which is in turn connected to both a Transec 206, a transmission security component that enables switching between radio frequencies to prevent electronic eavesdropping, and a delay/ multi- band radio 208, which is connected to a modem 210, which is connected to a Comsec 212, a communications security component, such as an encryption unit, which is linked to both the Transec 206 and to a decoder 214.

[0079] Figure 4 is a flow chart of one embodiment of the present invention and its relationship to components and client processors. The tasking engine 110 assigns tasks to the framework engine 100, an element of which is a configuration / reconfiguration engine 104. The configuration engine 104 then assigns elements of the plan to processing managers 304 associated with components 306.

[0080] Figure 5 is a block diagram of the configuration manager according to one embodiment of the present invention. The tasking engine 110 provides a dynamic plan to the matching comparing means 404 that compares that plan with the resources available to it, as listed in the current system configuration 106 with pretested templates stored in a library 108. The matching comparing means then assigns the task to available processors in such a way as to minimize demands on the system.

[0081] Fig. 6 is a flow chart illustrating the steps in the configuration method according to one embodiment of the present invention. A plan or schematic 102 is matched with at least one template having the same or similar system requirements. The results of this match are then compared with the current configuration 106 to see is the necessary processors are available to execute the chosen templates. The best templates are those that minimize demands on system resources. The optimal solution 602 is then used to configure the system and assign tasks to processors.

[0082] One embodiment of the present invention system for management of a distributed computing system including a tasking manager, whereby a plan is formulated describing a task to be carried out by the distributed computing system and detailing resources required for execution of the task a framework manager, communicating with the tasking manager and comprising at least one configuration manager, including a library within which are disposed at least one configuration template a means for ascertaining the current configuration of the distributed computing system a component assignment unit communicating with said tasking manager, the library, and the means for ascertaining the current configuration; and an output whereby the configuration manager directs the allocation of tasks to available system resources.

[0083] In such a system for management of a distributed computing system the configuration template may detail a tested configuration compliant with the distributed computing system and the component assignment unit may be configured to: compare said required resources with said each configuration template in said library to select usable configuration templates; compare said usable configuration templates with said means for ascertaining the current configuration; determine if system resources required by said usable templates are available; and select, based on comparisons of said usable templates with said current configuration, an optimal resource allocation for execution of said task.

[0084] According to one embodiment, system resources may include software processing units and hardware components. Such system resources may be adapted to RF applications such as tuners, channelizers, transmission security components, delay/multi-band radios, modems, communication security components, and decoders.

[0085] One embodiment of the present invention is a configuration manager for the assignment of task elements to processing units in a component based distributed system, including a library wherein are disposed at least one configuration template; a means for ascertaining currently available system resources of the distributed computing system; a component assignment unit communicating with a tasking manager, the library, and the means for ascertaining currently available system resources; and means for communicating assignments of the task elements to the processing units.

[0086] In this embodiment, the component assignment unit may be configured to: compare said required resources with said each configuration template in said library to select usable configuration templates; compare the usable configuration templates with the means for ascertaining the current configuration; determine if system resources required by the usable templates are available; and select, based on comparisons of the usable templates with the current configuration, an optimal resource allocation for execution of the task.

[0087] According to one embodiment, system resources may include software processing units or hardware components. Such system resources may be adapted to RF applications, such as tuners, channelizers, transmission security components, delay/multi-band radios, modems, communication security components, and decoders.

[0088] One embodiment of the present invention is a method for the configuration of components in a component based distributed system including the steps of: obtaining a plan schematic; obtaining at least one configuration template from a configuration template library; matching the configuration template with system requirements defined by the plan schematic; generating a preliminary result from the step of comparing the system requirements with the configuration template; obtaining a current configuration of components in the component based distributed system describing currently available system resources; matching the current configuration with the preliminary result; generating a final result of the step of matching the current configuration and the preliminary result; and communicating the result to distributed processing centers.

[0089] According to one embodiment, configuration templates include tested configurations for the component based distributed system. The preliminary result may include a plurality of useable templates while the final result is an optimal configuration for the components given said current configuration of the system.

[0090] According to one embodiment of the present invention, the method is executed within a framework manager. A tasking engine may formulate the plan.

[0091] According to one embodiment of the invention, matching the current configuration with the preliminary result is undertaken so as to obtain a template with an optimal correlation between the requirements of the template and resources available to the system.

[0092] According to one embodiment of the present invention, components may include software processing units or hardware components. Such components may be adapted to RF applications.

[0093] Included herein as Appendix A, are portions of source code that describe one embodiment of the invention. Those skilled in the art readily appreciate that the functionality expressed therein is not limited by the literal expression of the code and that there are numerous equivalent manners to present the functions. The comments and source code depict a single embodiment of the present invention. [0094] Numerous characteristics and advantages have been set forth in the foregoing description, together with details of structures and functions of a configuration engine, and the novel features thereof are pointed out in appended claims. The disclosure, however, is illustrative only, and changes may be made in arrangement and details, within the principle of the invention, to the full extent indicated by the broad general meaning of the terms in which the appended claims are expressed.

[0095] The objects and advantages of the invention may be further realized and attained by means of the instrumentalities and combinations particularly pointed out in the appended claims. Accordingly, the drawing and description are to be regarded as illustrative in nature, and not as restrictive.

APPENDIX A

* Summary Configuration Manager IDL Definitions (flug-in) to the

^* FrameworkManager

Copyright 2001 BAE SYSTEMS IEWS

^* [legend]

#ifndef ConfigurationManagerjdl #define ConfigurationManagerjdl

#include "Utility.idl"

#include "FrameworkManager. id I"

//module Framework {

// Specification for a single component type struct ComponentTemplateRcd { string componentName; long maxlnstances; long maxTransformsPerlnstance;

}; typedef sequence<ComponentTemplateRcd> ComponentTemplateList;

// Specification for a specific processor configuration struct ProcessorTemplateRcd { string templateName;

ComponentTemplateList components;

}; typedef sequence<ProcessorTemplateRcd> ProcessorTemplateList;

// An actual processor record struct ProcessorResourceRcd { string hostName; boolean loaded;

ProcessorTemplateRcd currentTemplate; ProcessorManager theProcessorManager;

}; typedef sequence<ProcessorResourceRcd> ProcessorResourceList;

// Mapping of templates to processors (legal combinations) struct TemplateToProcessorMappingRcd { string hostMatchString; // host name (with ^* as wildcards)

StringList allowedTemplates;

}; typedef sequence<TemplateToProcessorMappingRcd> TemplateToProcessorMappingList;

// configuration database; template definitions and processor mappings struct ConfigurationDatabase {

ProcessorTemplateList templates;

TemplateToProcessorMappingList mappings;

};

// Interfavce to manager processor configurations interface ConfigurationManager {

// get the set of possible processor configurations ConfigurationDatabase getConfigurationDatabaseO raises (unknownServerException);

// set the list of possible processor configurations void setConfigurationDatabase(in ConfigurationDatabase db) raises (unknownServerException, ParameterError);

// Save the list to a text file void saveConfigurationDatabase(in string filespec) raises (unknownServerException, ParameterError);

// restore the list to a text file void restoreConfigurationDatabase(in string filespec) raises (unknownServerException, ParameterError);

// Get the current reesources and assocaited template for each processor ProcessorResourceList getCurrentResources() raises (unknownServerException);

// determine if enough resources exist for a plan to be loaded boolean isPlanConfιgurable(in FrameworkManagerPlan plan) raises (unknownServerException, ParameterError);

// Configure a plan; assiging components to processors and connecting them // return of true indicates sucess boolean configurePlan(inout FrameworkManagerPlan plan) raises (unknownServerException, ParameterError);

};

// };

#endif

/* ft*****_***_*********************************************_******_**************_***

^* Summary Tasking Manager Server<p>

* Other Info The configuration manager serves as a utility class the the

* tasking manager may use to automatically assign components

* to processors given a specified plan. It currently will not

* share processors with multiple plans, however the hooks are

* there to implement with multiple plans, however the hooks

* are there to implement

Copyright 2001 BAE SYSTEMS IEWS

* [legend]

* iππ A *********************************************************** package Framework; import Java. util.^*; import java. io.^*;

/*_*

^* Configuration manager plug-in class. This class assigns components to processors

^* based on templates that are read from configuration files. The template file contains

^* how many of each component will fit on a processor. The file is organized as a

* series of configuration templates. A template represenets a known working configuration.

^* This file also contains which templates may be run on which processors by name. ^*/ public class ConfigurationManagerlmpl extends ConfigurationManagerPOA {

/^** Grace period given for a component to complete registration */ private static final int recentPlanTimeout = 6000; // ms

/^** Plan name for an unasigned component ^*/ public static final String unasignedProcessorName = "unassigned";

/** navigation back to the framewo. k manager private FrameworkManager theFrameworkManager = null;

/^** database built from the configuration file private Framework.ConfigurationDatabase theConfigurationDb;

/** current processor configurations; discovered from the framework manager 7 private ProcessorConfigurationRcdQ theConfiguration;

/_**

* Extra output to log. set by ConfigurationManagerDisplay

* option in the TaskingServer.properties file 7

public boolean debugDisplay = false; ^* a record of recently allocated plan to cover the condition where

^* a plan has been allocated but all the components have not loaded and

^* started. 7

ArrayList recentPlans = new ArrayList();

/**

* Processor list of best processors to deploy for a given template

* The natural order is having the largest score first. 7 private static class BestProcessorRcd implements Comparable { int score; int processorlndex;

BestProcessorRcd(int score.int processorlndex) { this.score = score; this, processorlndex = processorlndex; }

public int compareTo(Object other) { BestProcessorRcd otherl = (BestProcessorRcd)other; if (otherl .score > score) { return 1 ;

} else if (otherl .score < score) { return -1 ;

} return 0;

}

}

/_**

* Constructor expected to be called by the tasking manager 7 public ConfigurationManagerlmpl(FrameworkManager fm) { theFrameworkManager = fm; String str = TaskingServer.launchProperties.getProperty("ConfigurationManagerDisplay"); if ((str != null) && (str.equals("true"))) { debugDisplay = true; }

try { restoreConfigurationDatabase(TaskingServer.configDBFile); if (fm != null) { buildTheLoadedConfiguration(true,null);

} } catch (Exception ex) {

Log.debugException(ex); } }

**

* Post construction registration of the framework manager 7 public void registerFrameworkManager(FrameworkManager fm) { theFrameworkManager = fm; try . if (fm != null) { buildTheLoadedConfiguration(true,null); } } catch (Exception ex) { Log.debugException(ex); } }

/_**

* return if a processor name is unassigned or null 7 public static boolean isUnassigned(String name) { if (name == null) { return true; }

String lowerName = name.toLowerCase(); if (lowerName.equals(unasignedProcessorName)) { return true;

} else { return false; } }

/_**

* Return true if a plan is completely assigned 7 public static boolean allAssigned(FrameworkManagerPlan plan) {

// have we finished boolean ret = true; for (int j=0;j<plan.components.length;j++) { if (isUnassigned(plan.components[j].processorName)) { ret = false; break; } }

return ret; }

/_**

* Record of a recently allocated plans 7 class recentPlanRcd { long allocationTime; FrameworkManagerPlan thePlan;

recentPlanRcd(FrameworkManagerPlan plan) { allocationTime = System. currentTimeMillis(); thePlan = plan; } }

* Cleanup recent plan records 7 private void cleanupRecentPlans() { for (int k=0;k<recentPlans.size();k++) { recentPlanRcd rp = (recentPlanRcd)recentPlans.get(k); if ((System. currentTimeMillisO - rp.allocationTime) > recentPlanTimeout) { recentPlans.remove(k); --; } } }

#_**

^* get a list of components in the recent plan list for a selected processor 7

StringO listOfRecentComponents(String processorName) {

// iterate all the plans matching processor names ArrayList ret = new ArrayList(); for (int k=0;k<recentPlans.size();k++) { recentPlanRcd rp = (recentPlanRcd)recentPlans.get(k); for (int j=0;j<rp.thePlan.components.length;j++) { if (processorName. equals(rp.thePlan.components[j]. processorName)) { ret.add(rp.thePlan.components[j].componentName); } } }

StringQ retList = new String[ret.size()]; retList = (StringQ)ret.toArray(retList); return retList; }

r * Return true if a processor is in the list of recent plans

7 boolean componentlnRecentPlans(int componentld) {

// scan all the plans looking for components for (int k=0;k<recentPlans.size();k++) { recentPlanRcd rp = (recentPlanRcd)recentPlans.get(k); for (int j=0;j<rp.thePlan.components.length;j++) { if (componentld == rp.thePlan.components[j]. componentld) { return true; } } }

// made it to here; then it is not there return false; }

_**

^* Local class to save the current number of components for each type 7 class componentStatusRcd {

String componentName; int nlnstances;

componentStatusRcd(String name) { componentName = name; nlnstances = 0; }

}

r

* Local class to save the configuration for a processor 7 class ProcessorConfigurationRcd {

ProcessorManager pm;

ComponentlnterfaceQ ciList;

TreeMap componentlnstances; // of type componentStatusRcd (used to count # instances)

String hostName;

StringQ availableComponents;

BitSet possibleTemplates; }

/_**

^* return true if a template may be run on a given processor 7 private boolean templateAllowed(String templateName.String processorName) { templateName = templateName.toLowerCase(); processorName = processorName.toLowerCase(); for (int k=0;k<theConfigurationDb.mappings.length;k++) { TemplateToProcessorMappingRcd map = theConfigurationDb.mappings[k]; if (hostNameMatch(map.hostMatchString, processorName)) { for (int j=0;j<map.allowedTemplates.length;j++) { if (templateName.equals(map.allowedTemplates[j])) { return true;

} } } } return false;

}

/*_*

* Determine if a processor may host a template given the mappings and the

* current number/type of components already loaded 7 private boolean templatePossible(ProcessorConfigurationRcd config, ProcessorTemplateRcd testEntry, FrameworkManagerPlan plan) {

// first see if it allowed boolean possible = templateAllowed(testEntry.templateName,config.hostName);

TreeMap unassignedComponents = null;

if (plan != null) { unassignedComponents = calculateComponentsToBeAllocated(plan); }

// if so; then look at the components on the processor and eliminate templates // not possible because the processor does not support all the components or // the capacity of a given type has been exceeded if (possible) { for (int k=0;k<testEntry.components.length;k++) { String name = testEntry.components[k].componentName; componentStatusRcd csr = (componentStatusRcd)config.componentlnstances.get(name); if (csr == null) { possible = false;

} else if (csr.nlnstances > testEntry.components[k].maxlnstances) { possible = false; } } }

// now make sure there are no extra components not in the template if (possible) {

// construct a binary tree of all components in the template

TreeSet ts = new TreeSet(); for (int k=0;k<testEntry.components.length;k++) { ts.add(testEntry.components[k].componentName); }

// test each component that has instances

Iterator it = config. componentlnstances. entrySet().iterator(); while (it.hasNext()) { Map.Entry entry = (Map.Entry)it.next(); componentStatusRcd csr = (componentStatusRcd)entry.getValue(); if ((csr.nlnstances > 0) && (!ts.contains(csr.componentName))) { possible = false; break; } } }

// verify there is still some room left on the processor if (possible) { boolean roomLeft = false; for (int k=0;k<testEntry.components.length;k++) { String name = testEntry.components[k].componentName; componentStatusRcd csr = (componentStatusRcd)config.componentlnstances.get(name); if ((unassignedComponents != null) && (lunassignedComponents.containsKey(name))) { csr = null;

} if ((csr != null) && (csr.nlnstances < testEntry.components[k].maxlnstances)) { roomLeft = true;

} } possible &= roomLeft;

}

return possible;

* For a given processer determine a set of templatess may still be used

7 private void setPossibleTemplates(ProcessorConfigurationRcd pc, FrameworkManagerPlan plan) {

if (debugDisplay) { Log.debugMessage("»» Recomputing Possible Templates for Host: " + pc.hostName); };

// Set a bit for all templates possible for a processor; event if partly // allocated pc.possibleTemplates = new BitSet(); for (int j=0;j<theConfigurationDb.templates.length;j++) { if (templatePossible(pc,theConfigurationDb.templates[j],plan)) { pc.possibleTemplates.set(j); if (debugDisplay) {

Log.debugMessage(" Template!" + j + "]=" + theConfigurationDb.templates[j].templateName +

" is still possible for processor: " + pc.hostName); } } }

/_**

* Match a host name the match string is a name witn and option '*' to indicate

* wildcard matches 7 public static boolean hostNameMatch(String matchPattern, String hostName) { try { int wildcardChar = matchPattern. indexOf('^*'); if (wildcardChar >= 0) { hostName = hostName.substring(0,wildcardChar); matchPattern = matchPattern. substring(0,wildcardChar);

} matchPattern = matchPattern.toLowerCase(); hostName = hostName.toLowerCase();

boolean match = hostName.equals(matchPattern); return (match); } catch (Exception ex) { return false; } }

_**

^* Build the loaded configuration (as reported by the frameworkManager)

* This function sets the local context to reflect the current state of the

* systems as reported by the framework manager and its respective processor

^* manager's. 7 void buildTheLoadedConfiguration(boolean printCapabilities.FrameworkManagerPlan plan) throws unknownServerException {

// Pull the stuff from the framework try {

ProcessorManagerO pms = theFrameworkManager.getProcessors(); theConfiguration = new ProcessorConfigurationRcd[pms.length]; for (int k=0;k<pms.length;k++) {

// fill out the record using the public processorManager corba interfaces ProcessorConfigurationRcd pc = new ProcessorConfigurationRcd(); theConfιguration[k] = pc; pc.pm = pms[k]; pc.hostName = pc.pm. hostName(); pc.availableComponents = pc.pm. listAvailableComponents(); for (int j=0;j<pc.availableComponents.length;j++) { if ((printCapabilities) && (debugDisplay)) { Log.debugMessage("»» Processor " + pc.hostName + " May Run [" + pc.availableComponentsU] + "]");

} if (pc.availableComponentsU]. indexOf( ") >= 0) { int lastlndex = pc.availableComponents[j].lastlndexOf("/"); pc.availableComponents[j] = pc.availableComponentsU]. substring(lastlndex+1,pc.availableComponents[j].length());

} // construct a treemap of the current instances for each possible component pc.componentlnstances = new TreeMap(); for (int j=0;j<pc.availableComponents.length;j++) { pc.componentlnstances. put(pc.availableComponents[j], new componentStatusRcd(pc.availableComponents[j])); }

// populate the database with recent components cleanupRecentPlans();

StringQ recentComponents = listOfRecentComponents(pc.hostName); for (int j=0;j<recentComponents.length;j++) { componentStatusRcd csr = (componentStatusRcd)pc.componentlnstances.get(recentComponents[j]); if (csr != null) { csr.nlnstances++;

} }

// now fetch the rest from the processor checking for duplicates ComponentStatusRcdrj loadedComponents = pc.pm. listLoadedComponents(); pc.ciList = new Componentlnterface[loadedComponents.length]; for (int j=0;j<pc.ciList.length;j++) { pc.ciList[j] = loadedComponents[j].componentRef; String name = loadedComponents[j]. componentName; int id = loadedComponents[j]. componentld; if (!componentlnRecentPlans(id)) { componentStatusRcd csr = (componentStatusRcd)pc.componentlnstances.get(name); if (csr != null) { csr.nlnstances++; } } }

// Set a bit for all templates possible for a processor; event if partly // allocated setPossibleTemplates(pc,plan);

} } catch (Exception ex) {

Log.debugException(ex); throw ServerExceptionHelper.getUnknownServerException(ex); }

}

/_**

* Determine if a component type is supported (could be possibly loaded)

^* within the processor configuration 7 boolean componentSupported(String componentName) { for (int k=0;k<theConfiguration.length;k++) { for (int j=0;j<theConfiguration[k].availableComponents.length;j++) { if (componentName.equals(theConfiguration[k].availableComponents[j])) { return true; } } } return false;

}

/_**

* return true id a component is loaded on any processor 7 boolean componentLoaded(String componentName) { for (int k=0;k<theConfiguration.length;k++) { if (theConfiguration[k].componentlnstances.get(componentName) != null) { return true;

} } return false; Parse a template definition in the form:

* template <name> {

^* <component> : <maxlnstances> [: <maxTransforms>]

^* [;<component> : <maxlnstances {: <maxTransforms>]>] * }

* For example:

*

^* template feProcessorl {

^* channelizer : 1 ;

* mbr : 1 : 25;

* };

7 private ProcessorTemplateRcd parseTemplate(ParsedFileReader pfr) throws Framework.ParameterError {

ProcessorTemplateRcd ret = null;

try {

// read keyword and the template name String keyword = pfr.currentToken(); pfr.fetchNextToken();

String templateName = pfr.currentToken().toLowerCase(); pfr.fetchNextToken(); if (!pfr.currentToken().equals("{")) { throw pfr.parseError();

} else { // process each component record ArrayList components = new ArrayList(); pfr.fetchNextToken(); boolean done = false; while (Idone) {

String componentName = pfr.currentToken(); int maxlnstances = -1 ; int maxTransforms = 1 ; pfr.fetchNextToken();

String colon = pfr.currentToken(); if (!colon.equals(":")) { throw pfr.parseError();

} else { pfr.fetchNextToken(); maxlnstances = pfr.currentTokenAslnt(); pfr.fetchNextToken(); if (pfr.currentToken().equals(":")) { pfr.fetchNextToken(); maxTransforms = pfr.currentTokenAslnt(); pfr.fetchNextToken(); }

if (pfr.currentToken().equals(";")) { pfr.fetchNextToken();

} if (pfr.currentToken().equals("}")) { done = true; pfr.fetchNextToken(); if (pfr.currentToken().equals(";")) { pfr.fetchNextTokenO; } } }

ComponentTemplateRcd ctr = new ComponentTemplateRcd(componentName, maxlnstances, maxTransforms); components.add(ctr); }// not done

ComponentTemplateRcd[] componentList = new ComponentTemplateRcd[components.size()]; componentList = (ComponentTemplateRcd[])components.toArray(componentList); ret = new ProcessorTemplateRcd(templateName,componentList); }// start of body "{"

} catch (EOFException EOFex) {

Log.debugException(EOFex); throw new ParameterError("Parse Error read past EOF","","");

} return ret;

}

/_**

* Parse a processor mapping definition in the form

*

^* processor <hostname>[^*] {<template>[;<template>]};

_*

^* for exmaple:

*

* processor myri^* {feprocessor1,feprocessor2,feprocessor3,detectionProcessor} 7 private TemplateToProcessorMappingRcd parseTemplateToProcessorMapping(ParsedFileReader pfr) throws Framework. ParameterError {

TemplateToProcessorMappingRcd ret = null;

try {

// read keyword and the host

String keyword = pfr.currentToken(); pfr.fetchNextTokenO;

String hostnameMatchString = pfr.currentToken(); pfr.fetchNextTokenO; if (!pfr.currentToken().equals("{")) { throw pfr.parseErrorO;

} else {

// process each component record

ArrayList templateNameList = new ArrayList(); pfr.fetchNextTokenO; boolean done = false; while (Idone) {

String templateName = pfr.currentToken().toLowerCase(); templateNameList.add(templateName); pfr.fetchNextTokenO;

String delimiter = pfr.currentToken(); if (delimiter.equals(";")) { pfr.fetchNextTokenO; delimiter = pfr.currentToken();

} else if (delimiter.equalsf,")) { pfr.fetchNextTokenO; delimiter = pfr.currentToken();

} if (delimiter.equals("}")) { done = true; pfr.fetchNextTokenO; delimiter = pfr.currentToken(); if (delimiter.equals(";")) { pfr.fetchNextTokenO; } } }

StringQ list = new String[templateNameList.size()]; list = (StringD)templateNameList.toArray(list); ret = new TemplateToProcessorMappingRcd(hostnameMatchString,list); } } catch (EOFException EOFex) {

Log.debugException(EOFex); throw new ParameterErrorfParse Error read past EOF","","");

} return ret;

}

_**

^* IDL method to return the current configuration database 7 public Framework.ConfigurationDatabase getConfigurationDatabase () throws Framework.unknownServerException { return theConfigurationDb; }

/_**

* IDL method to establish the current configuration database 7 public void setConfigurationDatabase (Framework.ConfigurationDatabase db) throws Framework.unknownServerException,

Framework. ParameterError { theConfigurationDb = db; }

/_**

* IDL method to save the current configuration database as a file

* (TODO: not implemented) 7 public void saveConfigurationDatabase (Java. lang. String filespec) throws Framework.unknownServerException, Framework. ParameterError { throw new ParameterError("Not Implemented","",""); }

* IDL method to restrore a configuration database in text format from the

* specified file. 7 public void restoreConfigurationDatabase (Java. lang. String filespec) throws Framework.unknownServerException, Framework. ParameterError { try { ParsedFileReader pfr = new ParsedFileReader(new FilelnputStream(filespec)); ArrayList templateList = new ArrayList(); ArrayList processorList = new ArrayList(); while (pfr.isMoreTokensO) { String keyword = pfr.currentToken().toLowerCase(); if (keyword. equals("template")) { ProcessorTemplateRcd template = parseTemplate(pfr); templateList.add(template);

} else if (keyword. equals("processor")) {

TemplateToProcessorMappingRcd processor = parseTemplateToProcessorMapping(pfr); processorListadd(processor);

} else { throw pfr.parseError(); }

}

ProcessorTemplateRcdQ templates = new ProcessorTemplateRcd[templateList.size()]; templates = (ProcessorTemplateRcd[])templateList.toArray(templates); TemplateToProcessorMappingRcdO processors = new TemplateToProcessorMappingRcd[processorList.size()]; processors = (TemplateToProcessorMappingRcdQ)processorList.toArray(processors); theConfigurationDb = new ConfigurationDatabase(templates, processors);

} catch (lOException ex1) { theConfigurationDb = new ConfigurationDatabase(new ProcessorTemplateRcd[0], new TemplateToProcessorMappingRcd[0]); Log.debugException(ex1 ); throw new Framework. ParameterError("File Access Error", "Filespec",filespec);

} catch (Exception ex) { Log.debugException(ex); * IDL method to return the current configuration records.

^* (TODO: return templates; IDL change required for multimple templates) 7 public Framework.ProcessorResourceRcdQ getCurrentResources () throws Framework.unknownServerException {

Framework. ProcessorResourceRcdQ ret = new Framework.ProcessorResourceRcd[theConfiguration. length]; for (int k=0;k<ret.length;k++) { ret[k] = new ProcessorResourceRcd(theConfiguration[k]. hostName, true,null,theConfiguration[k].pm);

} return ret;

}

_**

* Calculate a template that contains all the unallocated components and their

^* required number of instances and save as a tree. The return is a tree of

^* ComponentTemplateRcd objects. 7 private TreeMap calculateComponentsToBeAllocated(Framework.FrameworkManagerPlan plan) {

// get a list of unassigned component names TreeMap unassigedComponents = new TreeMapO; ComponentPlanRcdO components = plan. components; for (int k=0;k<components.length;k++) { if (isUnassigned(components[k]. processorName)) { ComponentTemplateRcd ctr = (ComponentTemplateRcd)unassigedComponents.get(components[k]. componentName); if (ctr == null) { ctr = new ComponentTemplateRcd(components[k].componentName,0,0); unassigedComponents.put(components[k].componentName,ctr); } ctr.maxlnstances++; } } return unassigedComponents;

}

/*_*

^* Return a list of the best templates to satisfy a plan (least # of processors)

^* Note: the return is null if the plan may not be satisfied by the current list of

* templates 7 private ProcessorTemplateRcdO getBestTemplates(Framework. FrameworkManagerPlan plan) {

// make a tree of un allocated components

TreeMap unassignedComponents = calculateComponentsToBeAllocated(plan);

// Keep a list of templates to assign to processors ArrayList templatesToAssign = new ArrayList();

BitSet templatesThatCanStillWork = new BitSet(); for (int k=0;k<theConfiguration.length;k++) { templatesThatCanStillWork.or(theConfiguration[k].possibleTemplates); }

// mantain a score for each template indicating how many components it satisfies int scorerj = new int[theConfigurationDb.templates.length];

boolean assignmentsMade = true; while(assignmentsMade) {

// score the templates computing how may components may be satisfied on each assignmentsMade = false; for (int k=0;k<score.length;k++) {

ProcessorTemplateRcd pe = theConfigurationDb.templates[k]; score[k] = 0;

// only score templates that are viable and have not already been assigned if (templatesThatCanStillWork.get(k)) { for (int j=0;j<pe.components.length;j++) { ComponentTemplateRcd ctr = (ComponentTemplateRcd)unassignedComponents.get(pe.components[j].componentName); if (ctr != null) { int componentsThatCanBeAssigned = ctr.maxlnstances; if (pe.components[j]. maxlnstances < componentsThatCanBeAssigned) { componentsThatCanBeAssigned = pe.components[j]. maxlnstances;

} score[k] += componentsThatCanBeAssigned;

} } } }

// find the template that allocates the most components int maxScore = scorefO];

ProcessorTemplateRcd bestTemplate = theConfigurationDb. templates[0]; for (int k=1 ;k<score. length ;k++) { if (score[k] > maxScore) { bestTemplate = theConfigurationDb.templates[k]; maxScore = scorefk]; } }

// Make the assignment removing the components that are satisfied if (maxScore > 0) { for (int k=0;k<bestTemplate.components.length;k++) { ComponentTemplateRcd ctr = (ComponentTemplateRcd)unassignedComponents.get(bestTemplate.components[k].componentNa me); if (ctr != null) { assignmentsMade = true; ctr.maxlnstances -= bestTemplate.componentsfk]. maxlnstances; if (ctr.maxlnstances < 0) { ctr.maxlnstances = 0; } } } templatesToAssign.add(bestTemplate); for (int k=0;k<theConfigurationDb. templates. Iength;k++) { if (theConfigurationDb.templates[k].equals(bestTemplate)) { templatesThatCanStillWork.clear(k); break; } } }

}// while we are assigning components

// setup the return ProcessorTemplateRcdO ret ^{= nu|1}: if (templatesToAssign.size() > 0) { ret = new ProcessorTemplateRcd[templatesToAssign.size()]; templatesToAssign.toArray(ret); }

return ret;

^* return the spare capacity (# of component instances) for a processor

* given an input template 7 private int spareCapacity (ProcessorConfigurationRcd config, ProcessorTemplateRcd testEntry) { int spare = 0; for (int k=0;k<testEntry.components.length;k++) { String name = testEntry.components[k]. componentName; componentStatusRcd csr = (componentStatusRcd)config.componentlnstances.get(name); if (csr != null) { int instancesLeft = testEntry.componentsfk]. maxlnstances - csr.nlnstances; if (instancesLeft > 0) { spare += instancesLeft; } } } return spare;

}

* Return a processor that can handle a given template. This routine optionally

* will only use empty processors as specified by the useEmpty parameter.

* The return is the configuration database array index. (-1 indicates

^* failure).

* The return list is ordered such the items in the front of the list favor

^* processors that has the most extra capacity for loading components

^* for this template. 7 private intQ bestProcessors(ProcessorTemplateRcd thisTemplate, boolean useEmpty) {

// ordered list of BestProcessorRcd TreeSet ret = new TreeSet();

// find the index of the entry into the config database for thisTemplate int templatelndex = -1 ; for (int k = 0;k < theConfigurationDb.templates.length;k++) { if (theConfigurationDb.templates[k].equals(thisTemplate)) { templatelndex = k;

} }

// templates found; build the list processors scoring them in terms of extra // capacity if (templatelndex >= 0) {

// First look for an processor that may run all the components required // for this template for (int k=0;k<theConfiguration.length;k++) { ProcessorConfigurationRcd pc = theConfiguration[k];

boolean possible = pc.possibleTemplates.get(templatelndex); if (useEmpty) { possible = possible & (pc.ciList.length == 0); }

possible = possible & templateAllowed(thisTemplate.templateName,pc.hostName);

if (possible) { int score = spareCapacity(pc,thisTemplate); ret.add(new BestProcessorRcd(score,k)); } } }

// now load up the return intQ retArray = new int[ret.size()];

Iterator it = ret.iterator(); for (int k=0;k<retArray.length;k++) {

BestProcessorRcd bpr = (BestProcessorRcd)it.next(); retArray[k] = bpr. processorlndex;

} return retArray;

}

* Deploy the components associated with a plan given a specific template

* return true if any components have been assigned 7 private boolean deployComponents(Framework.FrameworkManagerPlan plan, int processorlndex, ProcessorTemplateRcd template) {

boolean anyDeployed = false; // return

ProcessorConfigurationRcd pc = theConfiguration[processorlndex];

Log.debugMessage(">»»Deploying components for processor: " + pc.hostName + " Plan: " ÷plan.name + " Template: " + template.templateName);

// iterate through the plan assigning components to this processor for (int k=0;k<plan.components.length;k++) { if (isUnassigned(plan.components[k]. processorName)) {

String componentName = plan. components[k]. componentName;

// locate the template red for this component

ComponentTemplateRcd ctr = null; for (int j=0;j<template.components.length;j++) { if (template.components[j]. componentName. equals(componentName)) { ctr = template.components[j];

} }

// now get the number of current assignments componentStatusRcd csr = (componentStatusRcd)pc.componentlnstances.get(componentName);

// both found; assigen the component if it will fit if ((csr != null) && (ctr != null)) { if (csr.nlnstances < ctr.maxlnstances) { csr.nlnstances++;

Log.debugMessagefAssigning Component " + plan. componentsfk]. componentName + "(" + plan.components[k].componentld +

") to Processor: " + pc.hostName); plan. componentsfk]. processorName = pc.hostName; anyDeployed = true; } } } }

// recompute the possible templates for this processor eliminating the ones // that have reached their allocation limit setPossibleTemplates(pcplan);

return anyDeployed;

/_**

^* determine if a plan is configured 7 private boolean isPlanConfigured(Framework. FrameworkManagerPlan plan) { boolean configured = true; for (int k=0;k<plan.components.length;k++) { if ((plan.components[k].theComponent == null) && isUnassigned(plan. components[k]. processorName)) { configured = false; } } return configured;

}

/_**

* IDL Method to return true if a plan may be configured. It does so by

* attempting to configure it. 7 public boolean isPlanConfigurable (Framework. FrameworkManagerPlan plan) throws Framework.unknownServerException, Framework. ParameterError { FrameworkManagerPlanHolder planHolder = new FrameworkManagerPlanHolder(plan); return configurePlan(planHolder); ^* Check if a configured plan is actually implementable 7 public boolean isConfiguredPlanlmplementable(FrameworkManagerPlan plan) throws Framework.unknownServerException {

TreeMap pcmap = new TreeMapO; boolean ret = true;

buildTheLoadedConfiguration(false,plan); for (int k=0;k<theConfiguration.length;k++) { pcmap. put(theConfiguration[k].hostName,theConfiguration[k]);

} for (int k=0;k<plan.components.length;k++) {

String name = plan. components[k]. componentName; String processor = plan. components[k]. processorName;

ProcessorConfigurationRcd pc = (ProcessorConfigurationRcd)pcmap.get(processor); if (pc == null) { Log.debugMessagef Warning processor " + processor + " Not Found; Component:" + name); ret = false;

} else {

Object found = pc.componentlnstances. get(name); if (found == null) { Log.debugMessage(" Warning processor " + processor + " does not list component " + name + " as supported"); ret = false; }

} } return ret;

} ^* IDL method to configure a plan. This function atempts to assign all the component

^* records that are unassigned to a valid processor (hostName) given the current

^* set of assignments and configuration database rules. The return is true if the

^* plan is validly configured and the plan with the processorName filled in for each component

* record. If the configuration failes then the processor assignments are restored

* to the values before the call. 7 public boolean configurePlan (Framework.FrameworkManagerPlanHolder plan) throws Framework.unknownServerException, Framework.ParameterError {

String origAssignments = new String[plan.value.components. length]; for (int k=0;k<origAssignments.length;k++) { origAssignments[k] = plan.value.components[k]. processorName; }

try { Log.debugMessage("==========Configuring Plan: " + plan.value.name + "

============--====== );

// is there anything to do if (isPlanConfigured(plan.value)) { boolean willRunPlan = isConfiguredP-anlmplementable(plan. value);

Log.debugMessagefConfiguration Manager has a problem with this manually allocated Plan"); return true; }

buildTheLoadedConfiguration(false,plan.value);

ProcessorTemplateRcdO templates = getBestTemplates(plan.value);

// loop while we are making progress and we have not made all the assignments boolean allAssigned = allAssigned(plan.value); boolean anyAssigned = true; while ((templates != null) && (lallAssigned) && (anyAssigned)) {

// try to assign components to processors by picking the set by

// picking the processors that will consume the max number of components

// as possible anyAssigned = false; for (int k=0;k<templates.length;k++) {

// first look for an empty processor into bestProcessorlndexes = bestProcessors(templates[k],true); boolean deployedThisPass = false; if (bestProcessorlndexes.length > 0) { for (int j=0;(j<bestProcessorlndexes.length) && (!deployedThisPass);j++) { deployedThisPass = deployComponents(plan.value, bestProcessorlndexes[j],templates[k]); anyAssigned |= deployedThisPass; } } else { bestProcessorlndexes = bestProcessors(templates[k],false); if (bestProcessorlndexes.length > 0) { for (int j=0;(j<bestProcessorlndexes. length) && (!deployedThisPass);j++) { deployedThisPass = deployComponents(plan.value, bestProcessorlndexes[j],templates[k]); anyAssigned |= deployedThisPass; ^

} } }

// not all done; try a new strategy allAssigned = allAssigned(plan.value);

} // for each template

// try a new strategy if all the templates have not been assigned if (lallAssigned) { templates = getBestTemplates(plan.value); }

} // for each trial

// did we perform a partial assignment? ; if so back it off if (allAssigned) { Log.debugMessage("»»»» Alocation sucessfull for Plan: " + plan.value.name);

// save the configured plan for future configuration requests recentPlans.add(new recentPlanRcd(plan.value));

} else { Log.debugMessage("»»»» Alocation failed: Plan: "+ plan.value.name); for (int k=0;k<origAssignments.length;k++) { if (isUnassigned(plan.value.components[k]. processorName)) { String str = "»»»» Component " + plan.value.componentsfk]. componentName + "(" + plan.value.components[k]. componentld + ") not Loaded; "; if (componentSupported(plan.value.components[k]. componentName)) { if (componentLoaded(plan.value.components[k]. componentName)) { str += "processors are overallocated";

} else { str += "matching template not found";

} } else { str += "component is not supported by any processors";

} Log.debugMessage(str);

} plan.value.componentsfk]. processorName = origAssignments[k];

} } return allAssigned;

} catch (Exception ex) { Log.debugException("»»»» Alocation failed",ex); for (int k=0;k<origAssignments.length;k++) { plan.value.components[k]. processorName = origAssignmentsfk];

} throw ServerExceptionHelper.getUnknownServerException(ex);

}

}

^* test program

7 public static void main(StringQ args) {

ConfigurationManagerlmpl cm = new ConfigurationManagerlmpl(null); try { cm.restoreConfigurationDatabase("d:\\test\\sampleConfiguration.txt");

Log.debugMessage("Done Processing");

} catch (Exception ex) {

Log.debugException(ex); }

}

}