1. © 2012 IBM Corporation The zEnterprise Unified Resource Manager

2. © 2012 IBM Corporation 2 2 Agenda  IBM zEnterprise and The Unified Resource Manager –Objectives –Structure  The Unified Resource Manager Functional Overview –Overview of major functions –Monitoring and Management capabilities –Interfaces

3. © 2012 IBM Corporation 3 zEnterprise Unified Resource Manager  Unifies management of resources, extending IBM System z qualities of service end-to-end across workloads  zEnterprise firmware that provides platform, hardware and workload- aware resource management  Optimized to host large scale database, transaction, and mission critical applications  The Most efficient platform for Large-scale Linux consolidation  Capable of massive scale up zEnterprise BladeCenter Extension (zBX) IBM zEnterprise ™ 196 (z196)  Selected IBM POWER7 ™ blades and IBM System x ® Blades for tens of thousands of AIX ® and Linux applications  High performance optimizers and appliances to accelerate time to insight and reduce cost  Dedicated high performance private network zEnterprise: A system of systems that unifies IT for optimized service delivery

4. © 2012 IBM Corporation 4 zEnterprise System Z CPU, Memory and IO PR/SM Z CPU, Memory and IO SE Virtual Machine zOS Linux Virtual Machine zOS Virtual Machine zOS Virtual Machine zOS z Blade Extension ISS DP Cell DWA pHyp AIX pHyp AIX pHyp AIX xHyp Linux xHyp Linux xHyp Linux AMM z Blade Extension ISS DP Cell DWA pHyp AIX pHyp AIX pHyp AIX xHyp Linux xHyp Linux xHyp Linux AMM z Blade Extension pHyp AIX pHyp AIX pHyp AIX xHyp Linux xHyp Linux xHyp Windows Linux x86Power AMM Virtual Machine zOS Linux Virtual Machine zOS Linux AIX Data Power Workload Resource Groups: Dynamic Mgmt and Reporting HMC (with Unified Resource Management) Private Data Network (IEDN) zHyp

5. © 2012 IBM Corporation 5 zEnterprise Ensemble  A zEnterprise Ensemble is a collection of 1..8 zEnterprise Nodes managed as a single virtualized pool of server resources  A zEnterprise Node is a single zCEC with 0..4 racks with up to 1..2 blade centers per rack  A zEnterprise node can be a member of only a single ensemble  The ensemble is the management scope for the Unified Resource Manager  A primary / alternate pair of HMCs provide the management console for the ensemble –The alternative HMC takes over in case the primary fails

6. © 2012 IBM Corporation 6 zEnterprise Networking  Network Simplification with the zEnterprise Network: Intra Ensemble Data Network (IEDN) Single physical, ethernet network combines System z CPC using OSA and Blade Center extention (zBx) Provides the data network for communications in the ensemble Physical network is virtualized through the System z and blade hypervisors Physical and virtual networks are managed from a central point of management  Secure communications (between System z and zBx) Physical security Server access to physical and virtual networks is controlled by the Unified Resource Manager  High Availability Redundant network hardware  Unique System z QoS Isolated / dedicated equipment Special purpose dedicated data network & OSA-Express (no encryption required)

7. © 2012 IBM Corporation 7 7 Hardware Management  Integrated Hardware Management across all elements of the multi-platform environment –Operational Controls –Firmware Inventory, Update, and Service –Hardware and Firmware Problem Detection, Reporting, and Call Home, Field Guided Repair and Verify –Physical Hardware Configuration, Backup, and Restore –Integrated Discovery and Resource Inventory

8. © 2012 IBM Corporation 8 8 Network Virtualization  Access to the IEDN is managed through the Virtual Network Ensemble Management functions in the HMC. –Management of Virtual Networks. Virtual Servers must belong to a Virtual Network to communicate on the IEDN Provides network isolation of virtual servers not on same virtual network System z Private Physical Layer 2 Network HMC zEnterprise Node z/OS LPARs Virtual Network Linux guests on z/VM AIX guests Power blade Virtual Network Management zBX Windows guests x86 blade

9. © 2012 IBM Corporation 9 9 Storage  Virtual Servers see Virtual Disks  Virtual Disks may be either –1:1 mapped storage resources –or for z/VM a minidisk created from a storage group  Storage Resources are –ECKD volumes (z/VM only) –FCP Logical Units (z/VM and IBM Blade hypervisors)  Manage Storage Access Lists –exists per hypervisor –defines accessible storage resources storage resources typically appear in different Storage Access Lists  Support Server and Storage Admin Roles: –Server Admin Identifies ensemble storage requiements Assign storage resources to hypervisors –Storage admin Assigns storage to the ensemble Configures SAN  z/OS Storage mgt approach unchanged Hypervisor Virtual Server VD1VD2VD3VS$ 1:1 mapped Virtual Disks storage group based Virtual Disks Storage Access List SR5 SR6 SR7 SR8 SR9SR3 SR4 SR1 SR2 SR4 SR5 SR7 SR8 SR9 (z/VM) Storage Group

10. © 2012 IBM Corporation 10 Hypervisor and Virtual Server Management  Hypervisor management –Start, stop, and query/list hypervisors –Update and repair an IBM Blade hypervisor –Monitor hypervisors and their resource use via Monitors Dashboard CPU Memory consumption  Virtual server life cycle management –Create / delete –Define CPU and memory –Attach to virtual networks –Attach to virtual disks –Attach virtual DVD images  Start / stop virtual servers –Support scheduled operations  Monitor virtual server resource usage through Monitors Dashboard  Move a virtual server definition to another hypervisor

11. © 2012 IBM Corporation 11 zEnterprise Performance Management  Platform management component responsible for goal-oriented resource monitoring, management, and reporting across the zEnterprise Ensemble –Core component responsible for definition and implementation of goal-oriented management policy –Workload monitoring and reporting based on management policy –Common approach to monitoring / management of platform resources across zEnterprise –Extend goal oriented approach of z/OS WLM to platform managed resources –Orchestration of autonomic management of resources across virtual servers Provide Intelligent Resource Director like function across the zEnterprise Management function will evolve over time –Pushes management directives to the SE, Hypervisors, and OS agents as required across the zEnterprise  Integration of HMC console support –Integrated UI for monitoring, display of workload topology relationships, status alerts, etc –Definition of Performance Management Goals and Policy Administration

12. © 2012 IBM Corporation 12 Platform Workload Resource Group  A Workload Resource Group is a grouping mechanism and “management view” of virtual servers supporting a business function  Provides the context within which associated platform resources are presented, monitored, reported, and managed  Management policies are associated to Workload Resource Groups –Performance Policy  Workload Resource Group definitions can be provided by higher level system management functions or created at the HMC Linux AIX z/OS AIX Workload = HR Workload = Payroll Ensemble Policy

13. © 2012 IBM Corporation 13 Workload Based Monitoring and Reporting  Provide reporting capability that shows usage of platform resources in a Workload context within a zEnterprise Ensemble scope –Across virtual servers / partitions supporting the Workload  Workload goal vs actual reporting  Drill down from overall Workload “performance health” view to contributions of individual virtual server –CPU, memory, network resource usage  Graphical views –Topology, trending graphs, etc  Links to system activity displays to show hardware utilization views  Reporting is limited to platform level resources, not trying to replicate tools that report on intra-OS resources and performance

14. © 2012 IBM Corporation 14 Network Monitors Dashboard  User Interface display that shows the following major views –Virtual Networks across IEDN, allowing drilling down to resources that comprise the virtual network: virtual switches, virtual servers, virtual network interfaces These views show the performance between the virtual switch and the guest Provide data “by” virtual network (i.e., by VLAN) and across all virtual networks –Physical Interfaces to the IEDN Virtual Switch Uplink interfaces to the physical interfaces that connect to the IEDN External TOR ports connected to a router –Physical Switches - TORs and ESMs  From the metrics collected, show –Performance statistics such as transfer rates over intervals –Both cumulative and interval metrics –Also includes raw metrics as collected –Resource state/status where available –Resource utilization relationships (e.g., Servers using an OSA)  Network Monitors Dashboard provides both current and historical displays

15. © 2012 IBM Corporation 15 Managing Resources across Virtual Servers  Manage processor resources across virtual servers to achieve workload goals –Detect that a virtual server is part of Workload not achieving goals –Determine that the virtual server performance can be improved with additional resources –Project impact on all effected Workloads of moving resources to virtual server –If good trade-off based on policy, redistribute processor resources Z CPU, Memory and IO SE PRSM Virtual Machine z/VM Virtual Machine MM Cell BladeData Power Blade Blade Center Xen Virtual Machine Xen Virtual Machine Xen Virtual Machine Cell Blade Data Power Blade MM Cell BladeData Power Blade Blade Center Xen Virtual Machine Xen Virtual Machine Xen Virtual Machine Cell Blade Data Power Blade MM Blade Center Hypervisor Virtual Machine Hypervisor Virtual Machine Hypervisor Virtual Machine Missing Goals!!! Virtual Machine HMC

16. © 2012 IBM Corporation 16 Enabling External Management Tools Provide API access to Unified Resource Manager function API allows programmatic access to the same underlying function as is accessed via the HMC UI Same resource types, instances and policy API functions corresponding to views and tasks in the UI Listing resource instances Creating, changing, deleting resource instances Operational control of resource instances Monitoring Etc. Enables management of an ensemble from external (to HMC) tools Key scenarios: Discovery, Monitoring and Provisioning use cases zEnterprise System zHMC UI API

17. © 2012 IBM Corporation 17 Tivoli monitoring can track and manage Workload Resource Groups across zEnterprise Business Service Management and SOA Add existing operating system, middleware, and application insights Monitoring Virtual Servers Storage Network Policy Workload Dashboard Events Unified Resource Manager APIs zBX Blade HW Resources System z PR/SM Z196/z114 z/OS System p Linux x86 AIX zEnterprise System Linux z/VM HW Resources PowerVM AIX Linux INxHY Data Power X150z

18. © 2012 IBM Corporation 18 Tivoli System Automation can ensure availability of Workload Resource Groups and Business Services Virtual Servers Storage Network Policy System Automation Integrated Workload availability for normal operations and disaster recovery! Business Service Workload Unified Resource Manager APIs Automate Operations zBX Blade HW Resources System z PR/SM Z196/z114 z/OS System p Linux x86 AIX zEnterprise System Linux z/VM HW Resources PowerVM AIX Linux INxHY Data Power X150z

19. © 2012 IBM Corporation 19 Unified Resource Manager Summary  Integrated Ensemble Management Firmware –Major virtualization and management functions are pre-integrated with the system, with less do-it-yourself planning, assembling, testing, … –System elements are developed, configured, and tested to work together compatibly –Integrate, virtualize, monitor, and manage platform resources in accordance with specified business workload objectives –Providing transparent value to all hosted software stacks –Built-in capability for upward integration with Tivoli Service Management