Managing Service Levels for Optimal Performance Availability by bxl82158

VIEWS: 12 PAGES: 45

									    <Insert Picture Here>




Managing Service Levels for Optimal Performance &
Availability
Rajiv Taori
Director, Product Management
Disclaimer

THE FOLLOWING IS INTENDED TO OUTLINE OUR
GENERAL PRODUCT DIRECTION. IT IS INTENDED
FOR INFORMATION PURPOSES ONLY, AND MAY
NOT BE INCORPORATED INTO ANY CONTRACT. IT
IS NOT A COMMITMENT TO DELIVER ANY
MATERIAL, CODE, OR FUNCTIONALITY, AND
SHOULD NOT BE RELIED UPON IN MAKING
PURCHASING DECISION. THE DEVELOPMENT,
RELEASE, AND TIMING OF ANY FEATURES OR
FUNCTIONALITY DESCRIBED FOR ORACLE'S
PRODUCTS REMAINS AT THE SOLE DISCRETION
OF ORACLE.
    Oracle, The Information Company

                              Enterprise Manager
• Configuration Management    • Application Performance Mgmt      • Dashboards
• Service Level Management    • Lifecycle Management              • Heterogeneous Support


     Database                         Fusion                     Information Age
  Grid Computing                    Middleware                     Applications
• Database                    • Application Server             • E-Business Suite
• Real Application Clusters   • SOA / Integration              • PeopleSoft Enterprise
• Partitioning                • Business Intelligence          • Siebel

• OLAP and Data Mining        • Content Management             • JD Edwards
                                                               • Oracle Retail
• Security                    • Identity Management
                                                               • i-Flex
• Spatial                     • Data Hubs
                                                               • Communications Billing
• Oracle Lite                 • Collaboration Services
                                                               • Utilities
• Times Ten                   • Java Development Tools
                                                               • ProfitLogic
                                                               • G-Log
 IT Management Challenges


• Modern data center based on SOA and Grid infrastructure
  • Dynamic, distributed computing using heterogeneous systems
  • No longer simple, standalone systems
• Regulatory compliance is becoming more complex
  • Controls are reactive and manual, costly and error prone
  • Automated IT controls are non-standardized and immature
• Traditional tools cannot adequately address today’s needs
  • Silo’ed management approach
  • Focused on bottom-up infrastructure, not business activity
   Today’s Environments

Users
                      Web Servers   Application Servers   Databases
        Services


        Order Entry


         Invoicing
                                    Message Bus

        HR Query



          Email
                      Hosts
   Management with Oracle Enterprise
   Manager
Users
                      Web Servers   Application Servers   Databases
        Services


        Order Entry



                                    Message Bus




                      Hosts
Enterprise Manager
Breadth of Capabilities


                                                  Enterprise
                                                   Manager


                                           Application Management




                                                                                                                         Extensibility and Interoperability
  Dashboards and Reports




                                                                                     (Open Standards, Open Interfaces)
                                       Oracle Infrastructure Management

                                     Extended Infrastructure Management

                                                          Service      Application
                           Configuration    Lifecycle     System
                                                           Level      Performance
                           Management      Automation    Monitoring
                                                        Management    Management


                                                   CMDB
                                           (Discover & Reconcile)
Service Level Management
Align IT and business customers on service levels


                                      • Define & Model
                                          – Define services and systems
                                          – Set service level goals
                                          – Map business service to system
                                            resources
                                          – Visualize in topology views

                                      • Monitor
                                          – Measure service performance
                                            against goals
                                          – Availability, Performance and
                                            Usage

                                      • Report
                                          – Single view of status of each
                                            service
                                          – Out-of-box and custom reports
Service Level Management Challenges


1. How do I negotiate and specify my SLAs? Which
   applications and systems support which services?
    -   What is the relationship between my services and
        infrastructure?

2. Can I track progress against my SLAs? Can I quickly
   identify problems once they occur?

3. How to provide visibility to my stakeholders on
   Service Level performance?
Oracle’s Approach to SLM
   •   Define & Model
        – Focus on end-user perspective
        – Map business service to IT resources
                                                 Oracle’s Difference
   •   Monitor
                                                 1.   Real-time view
        – Availability, Performance based on:
                                                      of SLM,
             • System                                 enabling you to
             • Service Tests                          resolve issues
        – Usage                                       before service
        – Service Level Compliance                    level violations


   •   Report                                    2.   SLA based on
                                                      both, system
        – Centralized reporting and Dashboards        and end user
        – Out-of-box reports                          metrics
        – Custom reports using Information
          Publisher
Define & Model Services


Service
  • Generic (Network services, custom,
     other protocols)
  • Web Application
  • Application Specific
       – Forms Application (EM 10.2.0.3)
       – Siebel, E-Business Suite,…
  • Aggregate

System
  – Key components responsible for supporting a service
  Service Availability & Root Cause

Test Based
   Defined on user experience
   Service is UP if “key” tests
   execute successfully




System Based
• Defines the dependency map
• Service is UP if “key” components are available
  Service Performance

Service Performance is defined:

• Using response metrics collected from beacons
  executing Service Tests or Transactions

And / Or:

• Using performance metrics of components in its
  System
    Service Usage
    (Business KPI’s that relate to running of IT)

•    Measures user demand of the service
•    Based on metrics of System components that support the service:
       •   Select metric from specific component
            E.g. Usage of webmail = ‘Active sessions’ of HTTP Server
       •   Statistically combine metric from set of components
            E.g. Minimum (Queries Per Hour) between LDAP Server1 and LDAP
                Server2

Business Metrics (available with upcoming release)
• Integration with BAM metrics
• Facilitated by Data Exchange Hub
• Examples:
       •   # of loan applications received per hour
       •   Number of orders processed
  Service Level Compliance

• Service Level parameters specification:
    –   Business Days, Hours
    –   Availability Criteria
    –   Performance Criteria
    –   Business Criteria

• Calculation of Service Level over a time-period as:
    – % of time during business hours Service meets the
      availability and performance criteria

• Comparison of achieved Service Levels to
  requirements in your SLA
 Root Cause Analysis

• Determines possible causes for service failures
• Checks up/down status of:
    •   every key component in underlying System
    •   dependent sub-services
    •   selected system component metrics that have reached
        ‘critical’ threshold
• Impact analysis
• Support for external events (EMC SMARTS) for
  network failures
 Automatic Service & System Modeling

Challenge
• Initial setup of a system model is error prone and time
  consuming
• Keeping track of changes to model (as you scale out, etc.) is
  also error prone

New in GC 10.2.03
• Automatic system model of J2EE application deployed in Oracle
  AS
• Change history tracking of system membership
• Configuration change history tracking of entire system (not just
  individual components)
• Customization of system model allowed – by adding / removing
  individual components
    System Dashboard

•   One view of
    overall status of
    all key components

•   At a glance
    monitoring of all
    critical indicators

•   Customizable views:
    metrics, configurations, policies, alerts

•   User editable “comment” fields for problem management
    tracking
  Reporting: Services Dashboards & Service Level
  Drilldowns
‘One-stop’ management for your most critical operations




                                         One view for overall status
                                         of each application
                                         Monitoring of all critical
                                         indicators of application at
                                         a glance:
                                           - Availability, Performance, Usage
                                         Service Level Compliance
Application Performance Management
Identify and resolve end-user problems quickly
                         • Monitor from end-user perspective
                            – Test application from representative geographies
                            – Measure real user performance by domain, visitor, etc.
                            – Http(s) and 10.1.2 Forms real user monitoring

                         • End-to-end system monitoring for complete
              Network


                           application infrastructure
                            – Comprehensive event management
              URL/JSP




                                • Real-time and historical performance
                                • Recommendation on metrics, thresholds
               Servlet




                                • Automation of corrective actions
                            – J2EE Application monitoring (top reports, performance
                              analysis, workload distribution)
                EJB




                         • Diagnostics and Tuning
              JDBC/
               SQL




                            – Rapid problem identification (Interactive Transaction
Application                   Trace, Request Performance, Page performance)
                            – Root Cause Analysis (RCA)
                            – Problem remediation/tuning
      End User Monitoring Approaches
   User
Communities                                    Web Servers    Application   Databases
                                                               Servers




                      Internet Cloud




                                                Hosts



Service Test Monitoring                     Real End User Monitoring
• Proactively monitor end user experience   • Monitor ALL real end user transactions
• Baseline performance & trend over time    • No instrumentation, No maintenance
                                                               location
    Service Test Monitoring - Beacons
    Rich Coverage for Service Protocols

                          Any service
Representative             protocol
    Client

     NY                     DNS, FTP,
                          HTTP(s), Ping,         Related
                        ICMP Ping, IMAP,   System Components
   Tokyo
                            JDBC SQL
                          Timing, LDAP,
   Paris                  NNTP, Oracle
                        SQL Timing, POP,
                           Port Checker
     SF
                         (includes SSL),
                          SMTP, SOAP,
   Berlin
                            TNS Ping,
                        Web Transaction,
                             WEBDAV
     Rio
                         Custom Script
 Web Transaction support for Rich
 Internet Applications – Use Cases
1. Monitor an application built with web2.0 (DHTML,
   javascript) technologies
2. Monitor an application where session data is passed
   as part of http headers – dynamic URLs
3. Obfuscate sensitive information from recorded
   transactions to increase security
4. Automatically detect successful execution of
   transactions
Feature – Benefit Summary
      Key Features:                         Benefits:
 User action recording &     Monitoring of Rich Internet
 playback                    Applications – e.g. using javacript,
                             javascript redirects, DHTML
 Http header data            No manual changes to recorded
 parameterization            scripts for applications with dynamic
                             URLs
 URL parameter or POST       Reduce steps needed to mask
 data obfuscation            sensitive information
 Automatic success string    Validation of business logic in
 suggestion                  execution of transaction
 Automatic mode detection    Self tuning capability hides complexity.
 between user action / URL   Automatically optimizes for
 fetchlet                    performance and accuracy
User Actions Recording




              Username & Password
               Click Page to
         EnterWait for Logout Load
                Login Page
                 Click Login
         DHTML / User Actions Mode
         HTTP Request Simulation Mode
Forms Transaction Monitoring
Test Performance Details
End-User Performance Monitoring
Monitoring of All URLs and Forms 10.1.2 applications
accessed by All Users

                              URL response times, # of hits
                                    View performance by:
                                        Domain
                                        Region
                                        Visitor
                                        Web Server
                                    Web server load and response distribution

                              Forms 10.1.2
                                    Commit, Query, Runform, Callform,
                                    Newform, Openform
                                •   EM 10.1.0.3+ : Total time, Server time
                                •   EM 10.2+: Total time, Server time,
                                    Database time
   Web Application
Application Diagnostics

                     Network       URL/JSP       Servlet       EJB          JDBC/SQL




  internet
    internet




Diagnostic Steps
  1. Diagnose problem to infrastructure component
               – Isolate problem to network/server: Request Performance
               – Isolate problem to tier: Interactive transaction trace
  2. Middle Tier Diagnostics
               – Isolate problem within tier: J2EE application monitoring
               – Identify application problems: Call stack analysis
  3. Database Tier Diagnostics
               – Performance diagnostics: ADDM, AWR
  1. Diagnose Problem to Infrastructure Component


Isolate problem to network / server
Request Performance
                                      Isolate problem to tier
                                      Interactive Transaction Trace
 2. Middle Tier Diagnostics

J2EE Application Monitoring
Workload distribution and performance analysis


                                             Application Problems
                                             Call Stack Analysis
Lifecycle Management
Cost effectively manage “many as one”
                                                          • Manage many as one
                                                                •   Virtualized resource views
                                                                •   System dashboards, topology views
                                                                •   Monitoring Templates
                                                                •   Admin: Backup, Secure, Start/Stop, etc.


                                                          • Cloning, Imaging and Provisioning
                                                                •   Pre-tested software image libraries
                                                                •   Automated software image deployment
                                                                •   Intelligent cloning makes adjustments for IP address
                                                                    host name, etc.
                                                                •   Provisioning of RAC/CRS and AS cluster
     Overview                                                   •   Provisioning of bare metal OS
                Web Cache   HTTP Server
                                                          • Life-cycle automation of Oracle patch process
                                          OC4J
                                                                •   Direct connection to MetaLink
                                           Visual ID of         •   Proactive notifications
                                           Root Cause           •   Automated distribution –one to many
                                             Failure
                                                                •   Critical Patch Facility (security)
                                                                •   OS patching
Overview - wizards
Reduce time to provision and patch

 • Automating installation and configuration
   • Store pre-configured, certified images in software library
   • Automate software image deployment (app server, bare
     metal OS)
   • Extend cluster via cloning
 • Automating patch updates
   • Connect directly to Oracle MetaLink
   • Receive notifications proactively via Critical Patch
     Advisory
   • Automate patch deployment (Oracle software, OS)



                             Provision on New Server:
          EM                 OS, App Server, Patches
    Deployment Procedures

•   Step-by-step best practices for application server provisioning
    tasks
•   Fully customizable
•   Four deployment procedures for Oracle Application Server:
     •   10.1.3.1SOA
     •   10.1.3
     •   10.1.2.0.2 (myJ2EE Application)
     •   Patching Application Server

Benefit:
• Customize EM to conform with your company’s processes, not the
  other way around
• Reduce errors and lower cost, by automating 100s of steps &
  1000s of documentation pages
    Configuration Management
    Manage impact of all changes

                                  •   CMDB reports       • Discover
                                  •   CMDB Statistics      – All hardware & software configurations
   Software                       •   Policy standards     – Oracle inventory, including patch history
 Installations                    •   Audit Reports          and updates
                                  •   Decision support
          SW and HW                   for incident,      • Analyze
         Configurations               problem, change,     – Monitor and audit change
                                      and release          – Configuration comparisons against
                  Resource            management             reference, saved or live configurations
                 Relationships
                                                         • Model-based decision support
                            Certified                      – CMDB maintains inventory of all CI’s
                          Configurations                     and their relationship
                                                           – Impact analysis, updates, baseline, and
                                         Best                verification of change
                                       Practices
                                                         • Foundation for compliance and
                                                           change control
                                                           – Map to ITIL and CoBIT frameworks
Configuration Management Database (CMDB)
                                                           – Policy Management
                                                           – Security configuration policies
Configuration Management Tool?
   AS Config Pack

• Collect,
  search,
  track,
  compare,
  and report
  on all
  hardware
  and software
  installation configurations

• Maintain a central inventory of software configuration data
  for J2EE, OHS and Webcache components
Policy-based Management


  Policy          Out of box ‘best practices’
                  Search for policy violations
                    – Security
                    – Configuration
                    – Storage
                  Calculate compliance scores
                    – Reflecting degree of
                       compliance (0-100%)
                  Notifications based on non-
                  compliance
                  Closed-loop remediation
SOA Runtime Governance
Manage for agility       • Auto-discovery & Auto-
                           creation
                            – Discover design-time BPEL
                              process dependencies
                            – Model end-to-end
                              environment as service
                              and system

                         • BPEL process monitoring
                            – Monitor process state,
                              latency, throughput
                            – Monitor process errors
                            – Monitor web service
                              performance
                            – In-context drill down into
                              BPEL Control

                         • BPEL PM Monitoring
                            – Monitor integration server
                              status, performance
                            – Aggregate view of all
                              processes
 SOA Management Pack – Key
 Takeaways
1. Only management solution for Oracle SOA Suite
   & BPEL
2. Manage, monitor, and diagnose problems in
  •   BPEL processes
  •   Integration (BPEL PM) server
3. Integration with Oracle BAM a differentiator
4. Support for BPEL 10.1.3+, SOA Suite 10.1.3+, BAM
   10.1.2+
Service Level Management




     OrderBooking
   Composite service

      SOA Infrastructure
           service

          BPEL Process
        availability service

             Credit Check
             Web Service
Customer Examples
            SLM Pack:    Beacons to monitor performance. Publish reports to execs
            Diag Pack:   Monitoring, metric comparison, trending, alerting
            Config Pack: Comparisons, change tracking

            SLM Pack:       Beacons to monitor portal/retail app availability
            Diag Pack:      Performance monitoring
            Config pack:    Compliance with corporate standards
            Provisioning:   Applying CPU

            SLM Pack:       Beacons to monitor performance on mobile portal
            Diag Pack:      Diagnosing middle tier performance problems


            Diag Pack:   AS performance monitoring and diagnostics
            Config Pack: Keeping multi-farms configuration consistent and in
                         compliance with corporate standards

            SLM Pack:    Performance monitoring of 30+ applications
            Diag Pack:   Diagnosing performance problems in multi-tiered
                         environment
            Config Pack: Tracking history of configuration changes

								
To top