WIXLIB

Note: Information contained in this briefing cannot be used/released without approval of authorPerpetual Enterprise ManagementService (PEMS)for C2 SOA DeploymentsPresented by:Erik King, C4IT Business Unitkinge@saic.com10th International Command and Control Research andTechnology SymposiumJune 14, 2005McLean, VAFrom Science to Solutions SAIC Company Proprietary1

ObjectivesDiscussion of the DoD Problem Domain Perpetual Enterprise Management Service (PEMS)as a Potential Solution Technical Solution (approach) Anticipated Capability (output)IR&D Partners PEMS Applicability to NCES Additional Resources From Science to Solutions SAIC Company Proprietary2

DoD Problem Domain:Enterprise Management of SOA SystemsCurrent systems are transitioning to Service Oriented Architectures(SOA) and as a result are becoming far more distributed – Applications are distributed via web services– Hardware platforms are being distributed via “grid” technologies– Operating systems can span multiple hardware configurations– Networks are becoming more diverse as IP traffic is possible over radio, ground, and satellitebased infrastructures– Sensor numbers are increasing as cost of collecting sensory data is reduced (ex. RFID)Understanding “Where”, “What”, and “How” a full system, andmission threads within that system reside, are comprised of, and operateis extremely complex in an SOA environment – With at least 3 dimensions of distribution (App, HW/OS, Network) it is difficult to pin downwhere the active logic of a user experience exists at any given time– Defining a “mission thread” beyond just its application components is critical if we are to haveany hope of managing these enterprise systemsNew methods of managing these distributed systems is required if trueService Level Objectives and Agreements are to be defined andmeasured. From Science to Solutions SAIC Company Proprietary3

Enterprise Management of SOA SystemsWhere does the management need to take place Mission thread or “userexperience” of a givencapability will occur acrossassets at all three layersServices/BusinessLogicsoftware engineerPlatformLayersystem administratorSDCONSOLEFDX 100BRIS/TLNKDSUCPUS3LPOKB1B2LOOPWIC 0 OKAUXWIC 1 OKCisco 1720NetworkLayernetwork administratorSensorGridFrom Science to Solutions SAIC Company Proprietary Each has its own potentialdistributed (and complex)architecture Each traditionally has itsown management tools andpersonnel who monitor healthindependent of one anotherNeed to provide tools andcapability to combine theseactivities into one in order tomanage end-to-end missionthread quality &performance4

Enterprise Management of SOA SystemsHow do we tackle the problem?IntelligentSensingSense &AssistSense &Respond Model theapplication,platform, andnetwork elementsof system/missionthreads via a realtime engine Provide tools tosystem administratorsto diagnose and solvecomplex issuesacross application,network, and systemboundaries Developknowledgebase thatmaps sensor data todiagnosis/resolutioninformation in order toestablish cases andpatterns Correlatemeaningful dataabout performancewithin thesethreads collect diagnoses fortrend and frequencyanalysis Provide real timefeedback to operators(both system adminsand potentially users)to suggest possiblecauses and resolutionto problems thatsurface Display threadmetrics in ameaningful way tohelp administratorspin-point problemsAutomatedSense &Respond Augmentknowledgebase withreal-time componentsthat automate fixes atapplication/interface,platform, and ornetwork level basedon pattern recognitionan case libraries Provide predictiveanalysis to operatorsbased on currentperformancecharacteristics ofmission threads andcorrespondinginfrastructurePerpetual Enterprise Management Service (PEMS)From Science to Solutions SAIC Company Proprietary5

SAIC Research Effort:Perpetual Enterprise Management Service (PEMS) Area of Focus:– Web Service Management (WSM) & Enterprise Systems Management (ESM) asit applies to the Net-Centric Command and Control domain. Current State of the Market:– COTS products exist in 3 distinct spaces: 1) Application or Business LogicLayer, 2) Platform/OS Layer, and 3) Network Layer. Vendors in each space arebeginning to provide web service and API hooks into their respectiveproducts. No single product exists that bridges these three areas to providereal time analysis and event management to promote self-healing ofdistributed SOA-based applications to improve the user experience. PEMS Objective:– Develop an integrated set of Web Services & UI capability that bridgecommercial products that serve the Business Logic Layer (ex. Amberpoint),Platform Layer (ex. IBM Tivoli) and Network Layer (ex. HP OpenView). Thiscapability would be extended to allow applications to conduct perpetualanalytics and subsequent system to system events to resolve applicationproblems (self heal) and improve user experience.From Science to Solutions SAIC Company Proprietary6

PEMS End-State VisionAPPGUIWebServicePerpetual Enterprise Management Service (PEMS)(correlates, fuses, aggregates in context to user on mentationTool(s)System ManagementNetwork ManagementSDCONSOLEBRIS/TFDX 100 LNKDSUCPUS3LPOKB1B2LOOPWIC 0 OKAUXWIC 1 OKCisco 1720From Science to Solutions SAIC Company Proprietary7

Technical Solution (1)COTS Product Classification (select some to help capture metrics)AmberPoint,Tealeaf,InfraVioBlue Titan,Actional,DataPower,Forum Systems,NetScaler,SantraApplication ManagementIBM Tivoli,Spectrum,CA Unicenter,BMC NMS (GOTS)System ManagementNetwork ManagementSDCONSOLEBRIS/TFDX 100 LNKDSUCPUS3LPOKB1B2LOOPWIC 0 OKAUXWIC 1 OKCisco 1720From Science to Solutions SAIC Company Proprietary8

Technical Solution (2)Identify business logic threads (“mission threads”) that define end-to-endfunctionality that needs to be monitoredFrom Science to Solutions SAIC Company Proprietary9

Technical Solution (3)Understand how that design is implemented from the users perspectiveFrom Science to Solutions SAIC Company Proprietary10

Technical Solution (3)Map out how business logic threads map to underlying Platform and NetworkLayer elements; identify appropriate metrics to collectBlue Force AggregationHelperAppMetricsRed Force StructureHelperRed Force StatusHelperBlue Force StructureHelperBlue Force By LocationHelperRed Force By LocationHelperGlobal Strike ECMRed Force CapabilitiesHelperRMIMission AnalysisCore Enterprise ServicesEJBCOA AnalysisWeb ServiceInformation Assurance /SecurityWorkspace ServiceSMD unable touse Xquery.Helper wrapsXQueryConfiguration WebServiceEnterprise ServiceManagementGCSS BridgeVirtually allservicesuse IA /SecurityESMpassivelymonitorsservices /does notexplicitlyconnectTheater IntegrationPlanning SystemBlue Force by LocationBridgeService DiscoveryEx. (Databaseaccess, webservice availability)Overlay ManagerHelperBlue Force AggregationBridgeBlue Force StructureBridgeMediationWeb Service(MarkLogicDocumentSer ice)PlatformMetricsEx. (CPU utilization,system availability,memory, HD space)NetworkMetricsPower Supply 0Power Supply 1SDCatalyst8500SER IESOKFDX100LNKDSUCPUS3LPB1WIC 0 OKB2LOOPAUXSwitchProcessorWIC 1 OKCisco 1720SDSTP1R CVECOM 3UTPCOM 4C OM 5N SRTW TLTTok en Ring 22TX10 BASE TRXC1Bay NetworksAUIPWAParSmiRps Adapter 1 OCIOFailFail Adapter 2PCMC IAERREt hernetBaySta ck A dvanced Remote NodeEt hernet 1SDSDSDCONSOLESTP1R CVECOM 3UTPCOM 4C OM 5N SRTW TLTTok en Ring 22FDX100LNKAUX10 BASE TRXC1Bay NetworksWIC 0 OKFDX 10 0 LNKAUXEt hernet 1AUIPWAParERREt hernetSmiRps Adapter 1 OCIOFailFail Adapter 2PCMC IABaySta ck A dvanced Remote NodeDSUCPUS3LPOKLOOPBRIS/TWIC 1 OKCisco 1720TXCONSOLEDSUCPUS3LPBRIS/ TB1WIC 0 OKB2OKB1LOOPB2Ex. (bandwidthlatency, LAN/WANavailability)CISCOSYSTEMSSDCONSOLEBRIS/ TW IC 1 OKCisco 1720From Science to Solutions SAIC Company Proprietary11

Technical Solution (4)NetworkPlatformBusiness LogicModel key threads using tools to identify underlying physical entitiesEthernetCISCO SYSTEMSFrom Science to Solutions SAIC Company 10x11x12x4x5x6xC7 8 9 101112A123456AB12

Technical Solution (5)Create an COTS-Integrated architecture for correlation, fusing, and reporting ofinstrumented data in a form that is useful to monitoring the ‘user experience’Scenario: User creates an ELINT ReportCESELINT Report Feed ELINT Report DataValidate User CredentialsDiscover ELINT Report ServiceValidate UserBusiness Processesimplemented in SOA fabricas web servicesCreate ReportSend Service Metric DataServices/Busine 2LOOPWIC 0 OKFDX 100 LNKAUXWIC 1 OKFCisco 1720NetworkLayerNIPRNet, SIPRNet,GIG-BE, JTRS,TSAT, (i.e., GIG)From Science to Solutions SAIC Company Proprietary13

Technical Solution (6)Implement the logical models into the commercial toolset and configure formanagement of a Common PictureFrom Science to Solutions SAIC Company Proprietary14

A Closer LookThe “Physical Model” of a key user threadFrom Science to Solutions SAIC Company Proprietary15

DemonstrationFrom Science to Solutions SAIC Company Proprietary16

Anticipated PEMS Capability / Output1. Real-time monitoring of 'end-to-end' mission threads / user experience(example: Blue Force Tracking by Location) that correlates the application(web service), platform, and network events that comprise the thread. Micromuse NetCool Suite is currently being used to visualize this correlated data. Web-based/portal-based interface will provide a network or systems engineer with adashboard which he/she can use to monitor important business processes(threads) When a problem occurs, the tool would help the operator pinpoint the root cause ofthe failure for remediation2. ESM (Health/Quality) Web Services that are available via a UDDI registrythat allows developers (and potential users of a service) to see how wellBPEL or SOA workflows (threads) are performing. Service would be discoverable like any other entity in the SOA fabric offering upthroughput, latency, and failure rate metrics (among others) for core and compositeservices that are being advertised in the UDDI registry Agencies being required to use NCES and other C2 services built on top of NCES(or other infrastructure) can understand the impact these services will have from aperformance perspective to their overall system designFrom Science to Solutions SAIC Company Proprietary17

Anticipated PEMS Capability / Output3. Knowledgebase / Decision Support Aide that helps operators improvetheir ability to recognize, predict, and automate resolutions to problemsthat are identified and remediated via #1 above. Leverages Case-base reasoning tools that Georgia Tech ResearchInstitute (GTRI) has built to capture the state of monitored events when aproblem occurs and the corresponding solution that is put in place tosolve these problems Over time, the knowledgebase will be able to provide predictive analysison 'what might occur' given a network/platform/application correlatedstate Will provide the system administrator or operator with a tool that wouldaide them in the remediation activity over time– System would collect problem parameters and corresponding remediation steps,but over time it would provide suggestions as real-time events are comparedagainst its pattern engine Ultimately, certain remediation activities would be able to be automated.From Science to Solutions SAIC Company Proprietary18

IRAD Proposed Partners Identified Partners: Micromuse Lucent Georgia Tech Research Institute (GTRI) Amberpoint IBM (Hardware and Tivoli support) Hewlett Packard (HP OpenView) Other Potential Partners: Actional BMC (Patrol)From Science to Solutions SAIC Company Proprietary19