160 likes | 364 Views
Chapter 2: Capacity. Chapter objectives. At the end of this chapter you should be able to describe: Mixed workloads SLA PR/SM and LPAR Parallel Sysplex Storage concepts and management Measuring capacity. Definition of Capacity. Capacity has several definitions:
E N D
Chapter objectives • At the end of this chapter you should be able to describe: • Mixed workloads • SLA • PR/SM and LPAR • Parallel Sysplex • Storage concepts and management • Measuring capacity
Definition of Capacity • Capacity has several definitions: • 1. The potential or suitability for holding, storing, or accommodating • 2. The facility or power to produce, perform, deploy or, simply process
Elements required for capacity (1/2) • Any computer system, from a single-user PC to a multi-thousand user IBM System z mainframe performs work in the following stages: • Load a program from a storage device, normally a disk, to central storage • Process the instructions in the program • Load any required data from disk or tape into central storage • Process the data and send the results to the desired end point
Elements required for capacity (2/2) • The relevant components to perform these steps and therefore for capacity estimation & planning are: • Processors • Central storage • Access input/output storage (for example disk or tape) • Also the network to these devices has to be considered
Many servers Few Servers • Many servers : • Data retention ? • Locking across servers ? • Connectivity ? • Few servers (IBM System z): • Large volumes of data: best in 1 place • No rewitting needed of applications • 1 total system
Mixed Workload • The approach to run mixed workload on one single system comes from the fact that in modern processor chips it is very unlikely, in a commercial environment, that one program could keep a processor fully utilized over a longer period of time. • For example, running jobs for • Online customers, • Online in-house users and • Batch work • on the same system.
Service Level Agreement • A Service Level Agreement (SLA) is an agreement between a service provider and a recipient, generally the server owner and a business unit. • SLAs are the baseline for capacity, availability and performance measurements and ratings. Example: • 95% of ATM transactions are completed in less than one second. • 90% of daily reports are completed by 6 A.M.
Managing the system to the SLA • Managing CPU • Managing Disk • Storage concepts (the address space) -> see next slide • Real storage mangement
Address space • The „Line“ 24 Bit systems • The „Bar“ 31 Bit systems
Architecture, running work & capacity • Program Status Word (PSW) • Interrupts • Machine check • Restart • Program Check • I/O • Supervisor call (SVC) • External • Supervisor call • Selection of work to run on a CP • Dispatching priority • CP utilization balance
Several servers on 1 physical machine • LPAR • Planning for downtime • Software • Hardware
Coupling Facility Z9 EC Z9 BC SYSPLEX Timer Z9 EC ESCON/FICON Z9 BC Shared Data IBM System z – Parallel Sysplex Parallel Sysplex
Measurements • Central Processor usage • Central storage usage • Access to disk storage
Summary • Capacity • Complex when you are : • In a mixed workload • On a large-scale commercial environment • Different workloads different requirements well defined set of SLAs needed • SLAs are also helpfull for: • Measurement of capacity bottlenecks • the planning of capacity upgrades.
24-bit 64-bit Address space Capacity Central storage CP Downtime I/O Interrupt LPAR Mixed Workload Parallel Sysplex PR/SM PSW Real Storage SLA SVC The Bar The Line Virtual Storage Key terms in this chapter