| |
| |
| Figure List | |
| |
| |
| Preface | |
| |
| |
| Introduction | |
| |
| |
| |
| Concepts and Architecture | |
| |
| |
| |
| Scaling Concepts | |
| |
| |
| What Is Scaling? | |
| |
| |
| Speedup | |
| |
| |
| Concurrency (Scaleup) | |
| |
| |
| Latches and Locks | |
| |
| |
| Why Lock? | |
| |
| |
| Things That Need Locking Latching | |
| |
| |
| Waiting on Latches and Locks | |
| |
| |
| Design Considerations to Reduce Lock Contention | |
| |
| |
| Linked Lists | |
| |
| |
| What Is a Linked List? | |
| |
| |
| What Are Linked Lists Used For? | |
| |
| |
| Optimizing Chain Lengths | |
| |
| |
| Hashing | |
| |
| |
| What Is Hashing? | |
| |
| |
| Caching | |
| |
| |
| Cache Fundamentals | |
| |
| |
| Memory Hierarchies | |
| |
| |
| Cache Reference Patterns | |
| |
| |
| I O Caching | |
| |
| |
| CPU Cache Concepts | |
| |
| |
| |
| |
| |
| Further Reading | |
| |
| |
| |
| Hardware Architectures and I O Subsystems | |
| |
| |
| Introduction to Hardware Architectures | |
| |
| |
| System Interconnects | |
| |
| |
| Bus Architecture | |
| |
| |
| Direct Memory Access (DMA) | |
| |
| |
| Cache Coherency | |
| |
| |
| Single Processor Architectures (Uniprocessors) | |
| |
| |
| Advantages | |
| |
| |
| Oracle on Uniprocessors | |
| |
| |
| Other Disadvantages | |
| |
| |
| Summary | |
| |
| |
| Symmetric Multiprocessors (SMPs) | |
| |
| |
| SMP Advantages | |
| |
| |
| Kernel Challenges | |
| |
| |
| Oracle on SMP Architectures | |
| |
| |
| Shared-Bus Limitations | |
| |
| |
| Summary | |
| |
| |
| Point-to-Point SMP | |
| |
| |
| Cache Coherency | |
| |
| |
| Summary | |
| |
| |
| Clustered SMP | |
| |
| |
| Clustering Types | |
| |
| |
| Summary | |
| |
| |
| Massively Parallel Processors (MPPs) | |
| |
| |
| Definition | |
| |
| |
| Oracle on MPP Systems | |
| |
| |
| Summary | |
| |
| |
| Cache Coherent Nonuniform Memory Access (ccNUMA) | |
| |
| |
| Definition | |
| |
| |
| Sequent NUMA-Q 20 | |
| |
| |
| SGI Origin 20 | |
| |
| |
| Oracle on NUMA Systems | |
| |
| |
| Summary | |
| |
| |
| Storage Systems | |
| |
| |
| I O Busses | |
| |
| |
| Controllers | |
| |
| |
| Disk Drives | |
| |
| |
| Disk Drive Sizing | |
| |
| |
| Redundancy | |
| |
| |
| RAID Levels | |
| |
| |
| RAID-5 Cached Disk Arrays: EMC Symmetrix | |
| |
| |
| |
| |
| |
| Further Reading | |
| |
| |
| |
| Building Support Software | |
| |
| |
| |
| Benchmark Concepts and Design | |
| |
| |
| Why Develop a Benchmark? | |
| |
| |
| Capacity Planning | |
| |
| |
| Upgrades | |
| |
| |
| Platform Changes | |
| |
| |
| What Are We Simulating? | |
| |
| |
| Defining Application Areas | |
| |
| |
| Defining the Application Split | |
| |
| |
| Careful with Those Assumptions, Eugene | |
| |
| |
| Selecting a Development Approach | |
| |
| |
| Using Remote Terminal Emulation Software | |
| |
| |
| Custom Simulation Development | |
| |
| |
| Building a Simulator Using RTE Software | |
| |
| |
| Building a Custom Benchmark Suite | |
| |
| |
| Programming Environment | |
| |
| |
| When the Application Has Not Been Written | |
| |
| |
| If the Application Exists: Trap and Emulate All SQL Calls | |
| |
| |
| Using Tracefiles to Generate a Simulation: An Introduction to dbaman | |
| |
| |
| Validate Server-Side System Utilization | |
| |
| |
| Building a Multiuser Framework | |
| |
| |
| Scale the Simulator | |
| |
| |
| Data Problems | |
| |
| |
| User Control Problems | |
| |
| |
| Simpler Methods for Use with dbaman | |
| |
| |
| Make It Easy to Run | |
| |
| |
| Portability | |
| |
| |
| Packaging | |
| |
| |
| Start-Up Scripts | |
| |
| |
| Automating Information Retrieval at Closedown | |
| |
| |
| Define Limitations in Advance | |
| |
| |
| A Benchmark Is Never Perfect | |
| |
| |
| Measure the Impact of the Inaccuracies | |
| |
| |
| |
| |
| |
| Further Reading | |
| |
| |
| |
| System Database Monitoring | |
| |
| |
| Why Monitor? | |
| |
| |
| Proactive Fault Detection | |
| |
| |
| Rapi | |