Cisco Unified Computing System (UCS) (Data Center): A Complete Reference Guide to the Cisco Data Center Virtualization Server Architecture

by Silvano Gai, Roger Andersson, Tommi Salli

Overview

The definitive guide to UCS and the Cisco^® Data Center Server: planning, architecture, components, deployment, and benefits

With its new Unified Computing System (UCS) family of products, Cisco has introduced a fundamentally new vision for data center computing: one that reduces ownership cost, improves agility, and radically simplifies management. In this book, three Cisco insiders thoroughly explain UCS, and offer practical insights for IT professionals and decision-makers who are evaluating or implementing it.

The authors establish the context for UCS by discussing the implications of virtualization, unified I/O, large memories and other key technologies, and showing how trends like cloud computing and green IT will drive the next-generation data center. Next, they take a closer look at the evolution of server CPU, memory, and I/O subsystems, covering advances such as the Intel^® XEON^® 5500, 5600, 7500, DDR3 memory, and unified I/O over 10 Gbps Ethernet.

Building on these fundamentals, the authors then discuss UCS in detail, showing how it systematically overcomes key limitations of current data center environments. They review UCS features, components, and architecture, and demonstrate how it can improve data center performance, reliability, simplicity, flexibility, and energy efficiency. Along the way, they offer realistic planning, installation, and migration guidance: everything decision-makers and technical implementers need to gain maximum value from UCS–now, and for years to come.

Silvano Gai has spent 11 years as Cisco Fellow, architecting Catalyst^®, MDS, and Nexus switches. He has written several books on networking, written multiple Internet Drafts and RFCs, and is responsible for 80 patents and applications. He teaches a course on this book’s topics at Stanford University.

Tommi Salli, Cisco Technical Marketing Engineer, has nearly 20 years of experience with servers and applications at Cisco, Sun, VERITAS, and Nuova Systems.

Roger Andersson, Cisco Manager, Technical Marketing, spent more than 12 years in the CLARiiON^® Engineering Division at EMC, and 5 years as Technical Product Manager at VERITAS/Symantec. He is now focused on Cisco UCS system management.

• Streamline data centers with UCS to systematically reduce cost of ownership
• Eliminate unnecessary server components–and their setup, management, power, cooling, and cabling
• Use UCS to scale service delivery, simplify service movement, and improve agility
• Review the latest advances in processor, memory, I/O, and virtualization architectures for data center servers
• Understand the specific technical advantages of UCS
• Integrate UCS 6100 Fabric Interconnect, Cisco UCS 2100 Series Fabric Extenders, UCS 5100 Series Blade Server Enclosures, UCS B-Series Blade Servers, UCS C-Series Rack Servers, and UCS Adapters
• Use Cisco UCS Manager to manage all Cisco UCS components as a single, seamless entity
• Integrate third-party management tools from companies like BMC ^®, CA ^®, EMC ^®, IBM ^®, Microsoft ^®, and VMware ^®
• Practice all this with a copy of Cisco Unified Computing System™ Platform Emulator Lite (UCSPE Lite) on the DVD in the back of the book

This book is part of the Networking Technology Series from Cisco Press®, which offers networking professionals valuable information for constructing efficient networks, understanding new technologies, and building successful careers.

Product Details

ISBN-13:	9781587141935
Publisher:	Pearson Education
Publication date:	06/01/2010
Series:	Networking Technology
Pages:	400
Product dimensions:	7.30(w) x 9.00(h) x 1.00(d)

About the Author

Silvano Gai, who grew up in a small village near Asti, Italy, has more than twenty-seven years of experience in computer engineering and computer networks. He is the author of several books and technical publications on computer networking as well as multiple Internet Drafts and RFCs. He is responsible for 30 issued patents and 50 patent applications. His background includes seven years as a full professor of Computer Engineering, tenure track, at Politecnico di Torino, Italy and seven years as a researcher at the CNR (Italian National Council for Scientific Research). For the past thirteen years, he has been in Silicon Valley where in the position of Cisco Fellow, he was an architect of the Cisco Catalyst family of network switches, of the Cisco MDS family of storage networking switches, and of the Nexus family of data center switches. Silvano teaches a course on the topics of this book at Stanford University.

Tommi Salli, who was born and raised in Finland, has close to 20 years of experience working with computers. He has extensive server and application background from companies like SUN Microsystems and VERITAS Software, which later got bought by Symantec from where he moved to Nuova Systems that got bought by Cisco. He has held different positions from Sales Engineer to Technology Scouting in the office of CTO from product management to architect and during his journey, he has been responsible for seven patent applications. He started his career in Finland, and for the past five years, he has been in Silicon Valley and is currently working for Cisco systems as a Technical Marketing Engineer.

Roger Andersson was born in Stockholm, Sweden. He has spent 20 years in the computer industry in both Sweden and the United States. Roger’s experience includes more than 12 years in the CLARiiON Engineering Division at EMC and five years at VERITAS/Symantec where Roger worked as a Technical Product Manager focusing on systems management, server, and application automated provisioning. Roger is currently working at Cisco as a Manager, Technical Marketing, where he is focused on the system management aspects of a Unified Computing System.

Table of Contents

Preface xvi

Nomenclature xvi

Chapter 1 Introduction 1

Data Center Challenges 1

Environmental Concerns–“Green” 2

Server Consolidation 3

Virtualization 4

Real Estate Power and Cooling 5

Cabling 5

Disaster Recovery 7

Network Virtualization 8

Desktop Virtualization 9

Cloud Computing 10

Evolution of Data Centers 10

Stand-Alone Servers 11

Scale-Up 12

Scale-Out 12

Scale-Up vs. Scale-Out 12

Rack-Optimized Servers 13

Blade Servers 14

Server Sprawl 15

Virtualization 17

Server Deployment Today 18

Unified Computing System (UCS) 18

Chapter 2 Server Architectures 23

The Processor Evolution 24

Sockets 24

Cores 25

Threads 27

Intel® Hyper-Threading Technology 27

Front-Side Bus 28

Dual Independent Buses 29

Dedicated High-Speed Interconnect 30

Intel® QuickPath Interconnect 31

The Memory Subsystem 33

SRAMs 34

DRAMs 34

SDRAMs 35

DIMMs 36

ECC and Chipkill® 38

Memory Ranks 39

UDIMMs and RDIMMs 40

DDR2 and DDR3 41

The I/O Subsystem 43

PCI Express® 43

Intel Microarchitectures 45

Platform Architecture 46

CPU Architecture 49

Virtualization Support 56

Advanced Reliability 59

Advanced Encryption Standard 60

Trusted Execution Technology 61

Chip Design 61

Chipset Virtualization Support 63

Intel® VT-d for Direct I/O 64

Intel® VT-c for Connectivity 65

VMDirectPath® 68

Chapter 3 UCS Enabling Technologies 69

Unified Fabric 69

10 Gigabit Ethernet 71

Lossless Ethernet 72

Terminology 72

PFC (Priority-Based Flow Control) 72

DCBX: Data Center Bridging eXchange 73

Bandwidth Management 74

FCoE (Fibre Channel over Ethernet) 75

Virtualization 81

Server Virtualization 81

SR-IOV 83

The IEEE Standard Effort 83

Port Extenders and Virtualization 84

VNTag 86

Fabric Extenders 88

VN-Link 90

Memory Expansion 93

Speed vs. Capacity 94

Capacity vs. Cost 94

How Much Memory Is Required? 95

NUMA 98

The UCS Approach 98

The UCS Advantage 101

Chapter 4 I/O Adapters 103

Disclaimer 104

The Intel® Approach 104

10 Gigabit Ethernet NIC Solutions 104

Intel® 82598 10 Gigabit Ethernet Controller (Oplin) 105

Support for Multi-Core CPUs 108

Hardware-Assisted Virtualization 109

Advanced Features for Storage over Ethernet 109

Intel® 82599 10 Gigabit Ethernet Controller (Niantic) 109

Improved Performance 112

Hardware-Assisted Virtualization 113

Support for DCB (Data Center Bridging) 114

Storage over Ethernet 115

Fibre Channel over Ethernet (FCoE) 115

Time Sync–IEEE 1588 115

Double VLAN 116

Security 116

Intel’s NetEffect™ iWARP Controller (NE020) 116

iWARP and RDMA 117

N2020 Architecture 119

Performance 120

Summary 122

Converged Network Adapters (CNAs) 123

Cisco® Palo 124

Emulex 129

Emulex OneConnect OCm10102-FC 130

FCoE Features 132

Ethernet Features 133

Functional Architecture 133

Deployment in UCS 133

Management of OneConnect UCNAs 134

Benefits of OneConnect UCNAs 136

QLogic® 137

8000 Series–First Generation CNA 137

8100 Series–Second Generation CNA 138

Broadcom® 144

BCM57711 Dual-Port 10GbE Controller 144

Advanced Integration 145

High-Performance Hardware Offload 146

Broadcom and UCS 151

Chapter 5 UCS B-Series Blade Servers 153

Components Overview 153

UCS Manager 154

UCS 6100 Series Fabric Interconnects 154

UCS 2100 Series Fabric Extenders 155

UCS 5100 Series Blade Server Chassis 156

UCS B-Series Blade Servers 157

I/O Adapters 159

Overall Organization 160

UCS C-Series Rack Servers 161

Detailed Description 161

UCS 6100 Series Fabric Interconnects 161

UCS 2104XP Fabric Extender 168

UCS 5108 Blade Server Chassis 172

Two-Socket Blade Architecture 172

UCS B200 Two-Socket Server 179

UCS B250 Extended Memory Server 181

Four-Socket Blade Architecture 182

UCS B440 Four-Socket Server 186

Description of Communication Flows 187

The Boot Sequences 187

Fabric Interconnect and UCSM 189

Fabric Extender 190

Baseboard Management Controller 190

Chapter 6 UCS C-Series Rack Servers 193

UCS C200 194

UCS C210 195

UCS C250 199

UCS C460 202

Processors 207

Adapters 212

Hard Disk 213

Management 213

Software 219

Physical Parameters 220

C200 220

C210 220

C250 220

C460 221

Weights 221

Chapter 7 UCS Manager 223

UCSM Overall Architecture 223

System Components 223

UCSM Is a Model-Driven Framework 227

Management Information Model 232

Available Integration Points 233

Interfaces 233

Standard (Cut-Through) Interfaces in a UCS 235

Standard Interfaces in a UCS 237

Native Interfaces in UCS 239

Operating Principles 240

Configuration Policies 242

Operational Policies 243

Global vs. Local Policies 244

Pools 244

Manual Population of Pools 250

Automatic Population of Pools 250

Service Profiles 252

Service Profile Templates 253

Organizations 255

Hierarchical Pool and Policy Resolution 256

Role-Based Access Control (RBAC) 257

Locales 258

Users and Authentication 259

UCSM and VMware’s vCenter Integration 259

Integration Architecture 260

Virtualization Support 260

Management Plane Integration 261

Port-Profiles 261

vNIC Template 262

Runtime Policy Resolution for Dynamic VIFs 262

UCS Manager and VM in GUI 264

Basic System Management with UCSM 264

Hardware Management 264

Example of a Chassis Discovery Process 265

Retirement of Hardware 266

Firmware Management 266

Firmware Download Formats 267

Firmware Life Cycle 268

Management Firmware Pack Policy 272

Host Firmware Pack Policy 272

The Stateless Computing Deployment Model 272

The Basic Computing Deployment Model 273

System Setup–Initial Setup 274

The Default Computing Deployment Model 274

The Stateless Computing Deployment Model 275

Requirements for Stateless Service Profiles 280

System Logging 287

Faults and Events 289

Audit Log 291

Backup and Restore of UCS Manager 292

Full State Backup 292

Configuration-Only Backup 292

Backing Up the UCS 292

Restoring a Configuration-Only Backup 293

Integrating with UCS 294

UCS Manager XML API 295

UCS XML API Object Naming 296

Method Categories 296

UCS Platform Emulator 301

Chapter 8 Third-Party Management Software 307

BMC® 307

Just-in-Time Provisioning 308

Embedded System Management 309

Business Service Provisioning 313

Composite Packaging 315

Configuration Management 316

Granular Access Control 318

Compliance 318

Vision for Automated and Efficient IT 319

CA® Management Integration with Cisco UCS 322

Integration Point 323

CA Infrastructure Management Integration 323

Discovery, Fault, and Service Modeling 324

Performance Management and Analytics 327

Automation 328

Change and Configuration Management 329

Service Profile and Application Templates 330

Automated Provisioning 331

Policy-Based Automation 332

User Self-Service 333

Private Cloud Deployments 333

EMC® Ionix Products for Cisco UCS 337

Unified Infrastructure Manager (UIM) 337

Data Center Insight (DCI) 341

IBM Tivoli Software Integration with Cisco UCS 344

Microsoft System Center 346

VMware vCenter 347

Communications 348

Configuration of the DVS 350

Virtual Machine Adapters 350

Resource Checks for DRS, HA, and FT 350

Chapter 9 Planning a UCS Blade Server Installation 353

The Owner of the UCS Domain 353

User Authentication 354

Power and Cooling 354

Physical Sizing and Environmental Requirements 357

Connectivity 359

Choosing the Right Cables 361

Twinax 361

Fiber 361

Bandwidth 362

OS and Application Support 363

Supported Storage Devices and Protocols 363

Planning for Redundancy 364

Power Supply Redundancy 364

I/O Redundancy 365

Ethernet Interfaces Redundancy 365

Fibre Channel Interfaces Redundancy 367

Bibliography 369

PCI Express 369

IEEE 802.3 369

Improvements to Ethernet 369

IEEE 802.1 Activities 369

FCoE 370

TRILL 370

Virtualization 370

Memory Subsystem 371

Intel® Processors 371

Data Centers 372

Green 372

Cloud Computing 372

Glossary 373

TOC, 9781587141935, 5/10/10