The Practical Performance Analyst

by Neil J Gunther M.SC., Ph.D., Raj Jain (Foreword by)

Overview

This book is the first to explain both theory and practical applications in an intuitive manner that computer professionals can understand and use to achieve higher systems performance.

Product Details

ISBN-13:	9780595126743
Publisher:	iUniverse, Incorporated
Publication date:	12/20/2000
Edition description:	New Edition
Pages:	468
Product dimensions:	8.28(w) x 11.02(h) x 0.98(d)

Table of Contents

	Foreword	xv
	Preface	xvii
	Acknowledgments	xxi
	Introduction	xxiii
Part 1	Foundations
Chapter 1.	About Time!	3
	What is Time?	4
	Physical Time	4
	Synchronization and Causality	5
	Discrete and Continuous Time	5
	Time Scales	6
	What is a Clock?	7
	Physical Clocks	7
	Distributed Physical Clocks	8
	Distributed Processing	8
	Binary Relations	9
	Logical Clocks	9
	Intervals and Events	11
	Virtual Clocks	11
	Response Time	12
	Gamma Distribution	13
	Server Response-time Distribution	15
	Network Response-time Distribution	16
	Timing Chains and Bottlenecks	16
	Travels with an SCSI Write Operation	19
	Travels with an NFS Read Operation	20
	A Time to Fall	22
	Availability	23
	Reliability and MTBF	24
	Distributed Systems	27
	Components in Series	27
	Components in Parallel	27
	Software Reliability	28
	Metastable Lifetimes	28
	Microscopic Metastability	28
	Macroscopic Metastability	31
	Summary	32
Chapter 2.	Queueing Theory for Those Who Can't Walt	35
	Terminology and Symbols	36
	Schematic Symbols	36
	Shorthand for Queues	38
	Observational Laws	41
	The Grocery Store: Checking It Out	42
	Queueing Analysis View	43
	Perceptions and Deceptions	44
	The Post Office: Snail Mail	44
	Little's Law Means a Lot	44
	A Visual Proof	46
	Server Utilization	51
	System Residency	51
	Infinite Source Queues	52
	A Simple Queue	52
	Twin Centers	57
	Parallel Centers	58
	Dual Server	60
	Multiserver Single Queue	61
	Erlang's Functions	63
	Feedback Centers	66
	Finite Source Queues	70
	Closed Queueing Center	70
	Dining Hall Analogy	71
	Finite Response Times	74
	Comparative Performance	75
	Approximating M/M/m/N/N by M/M/m	76
	Multiserver versus Uniserver	76
	Multiqueue versus Multiserver	78
	Non-Markov Queues	79
	Pollaczek-Khinchine Relation	80
	The Commuter Paradox	82
	Summary	82
	Summary of Queueing Formulas	83
Chapter 3.	Systems of Queues	85
	Queueing Circuits	86
	Open and Closed Circuits	86
	Poisson Properties	88
	M/M/1 Center	89
	M/M/m Center	89
	Jackson's Theorem	92
	Mean Value Analysis (MVA)	98
	The Arrival Theorem	98
	General Rules for Separability	104
	Service Policies	105
	Workload Classes	105
	Service-time Distributions	105
	State-dependent Service	106
	Arrival Processes	106
	Flow Balance	106
	One-step Behavior	106
	Device Homogeneity	106
	Modeling Computer Systems	107
	Some Traditional Examples	107
	Bounds and Bottlenecks	107
	Flow Equivalent and Load-dependent Servers	112
	Multiciass Workload Analysis	117
	What Queueing Circuits Can't Model (Well)	122
	Summary	124
Chapter 4.	Distributed Performance Management	125
	Networking Standards	127
	Some Preliminaries	128
	TCP versus UDP Service	128
	OSI Layers	129
	MIB Naming Tree	129
	System Group	130
	SNMP Standard	130
	SNMP	131
	MIB Extensions	132
	RMON	133
	RMON MIB	133
	Host MIB	134
	Enterprise MIB	135
	RDBMS MIB	135
	Limitations	137
	CMIP Standard	137
	System Management Hierarchy	138
	System Management Functions	139
	Workload Monitoring	139
	CMIP	140
	CMIP Groupings	141
	Limitations	141
	UMA Standard	142
	UMA Protocol	142
	UMA Data Format	144
	Limitations	146
	Functional Comparison	146
	Potential UMA-SNMP Integration	147
	Commercial Products	148
	Summary	150
Part 2	Applications
Chapter 5.	Commercial Parallelism	153
	Economies of Scale	153
	Hardware Parallelism	154
	Software Parallelism	157
	Parallel Platform Architectures	158
	Scientific Parallel Platforms	159
	Commercial Parallel Platforms	160
	Commercial Cluster Platforms	160
	Parallel Database Architectures	161
	Data Partitioning	163
	Informix XPS	165
	Oracle Parallel Server	165
	Parallel Query Scalability	168
	TPC-D Benchmarks	169
	Parallel Application Areas	170
	Information Warehousing	170
	Online Analytic Processing (OLAP)	170
	MPP Markets	170
	Summary	171
Chapter 6.	Parallel Systems	173
	Interconnect Networks	173
	Interconnect Topologies	173
	Interconnect Performance	175
	Speedup and Scaleup	177
	Speedup	178
	Speedup Bounds	181
	Scaleup	184
	Scaled Speedup	186
	Quadratic Scaleup	187
	Geometric Scaleup	188
	OLTP Scalability	188
	Super-seriality Phenomenon	189
	SMPs versus MPPs	192
	DSS Scalability	193
	Optimizing Query Performance	194
	Query Saturation Method	200
	Summary	204
Chapter 7.	Multiprocessor Systems	205
	Common SMP Architectures	205
	Cache Trashing	208
	SMP Performance Models	209
	Single-bus Models	210
	Processing Power	211
	Multiple-bus Models	212
	Cache Effects	216
	The Iron Law of Performance	226
	Quadratic Scaling Approximation	227
	Large-System Performance Reference Ratings	228
	External Throughput Rating	228
	Internal Throughput Rating	229
	Sizing by Numbers	231
	Gathering Public Data	232
	The Benchmark Workload	232
	Baseline Estimates	233
	Sizing Calculations	235
	Second-generation Sizing	236
	Current-generation Sizing	238
	Summary	239
Chapter 8.	Client-Server Applications	241
	What is Client/Server?	242
	Client/Server Is Just Pizza!	243
	Poor Pizza Time	244
	Probing Pizza Performance	246
	Capacity Planning	246
	The Enterprise Environment	246
	Workload Characterization	248
	Process Dependencies	249
	Spreadsheet Model	250
	PDQ Model	252
Scenario 1	Client Scaleup	263
Scenario 2	Gateway and Fileserver Upgrade	264
Scenario 3	DASD and LAN Upgrade	264
Scenario 4	LAN Upgrade Only	265
	Per Process Performance	266
	Three-Tiered Environments	266
	Summary	268
Chapter 9.	World Wide Web Servers	269
	What is the Web?	270
	The Scale of the Web	270
	Web Protocols	272
	Web Performance	273
	Web Server Capacity	276
	PDQ Models	277
	Fork-on-Demand Analysis	280
	Preforking Analysis	280
	Web Performance Tools	284
	Summary	284
Part 3	Innovations
Chapter 10.	Small Numbers, Big Consequences	289
	Small Chances	290
	Large Systems	291
	Virtual Memory	291
	Memory Stability	292
	Balancing Ball Bearings	295
	Birth-Death Model	297
	Larger Implications	299
	Adaptive Control	299
	Universality Classes	299
	Related Mathematical Techniques	300
	Summary	300
Chapter 11.	Paths, Potentials, and Probabilities	301
	Sample Paths and Potentials	302
	Probability Weighting	303
	Escape Routes	303
	The Tale Is in the Tail	305
	Getting a Piece of the Action	307
	MTTT Estimator	309
	The Path Integral Program	310
	Mathematical Particulars	311
	Calculus of Variations	311
	The Action	312
	Euler Equation	313
	Moment-Generating Function	314
	Minimal Action	314
	Summary	315
Chapter 12.	Large Transients in Packet-Switched Networks	317
	Packet Switch	317
	Buffer Dynamics	318
	Numerical Results	320
	ALOHA Stability	321
	Unstable Network	324
	Bistable Network	324
	Ethernet Stability	327
	Summary	328
Chapter 13.	Large Transients in Circuit-Switched Networks	329
	Network Transients	329
	Random Alternative Routing	330
	The Eriang Blocking Model	330
	Traffic Screening	331
	Linear Objective Function	333
	Mean Field Approximation	334
	Bilinear Model	336
	Bilinear Objective Function	337
	Stationary Points	338
	Critical Points and Phases	338
	Blocking Phases	339
	Summary	340
Chapter 14.	The Dynamics of Scaling	341
	One-Parameter Models	342
	Amdahl Scaling	342
	Geometric Scaling	343
	Quadratic Scaling	343
	Leveling the Playing Field	345
	Processing in Pictures	346
	Quadratic Scaling Diagrams	346
	Amdahl Scaling Diagrams	350
	Geometric Scaling Diagrams	352
	The Big Picture	353
	Rational Scaling Parameters	355
	Discrete Equations	355
	Stacking the Deck	358
	Overlap Functions	361
	Protocol Costs	363
	Summary	364
Part 4	Appendices
Appendix A.	PDQ User Guide	367
	Requirements	367
	Library Interface	368
	Data Types	368
	Global Variables	369
	Procedures	370
	From the B&N Reads Blog Page 1 of Related Subjects Computers Database Management Optimization/Performance - Databases Computers Database Management Optimization/Performance - Databases Customer Reviews