Numerics • COPYRIGHTED MATERIAL · overview, 146–147 performance issues, 165 for storage,...

• Numerics •32-bit servers, 175, 176–178, 181–18264-bit servers, 175, 178–180, 181–182

• A •administration. See IT; managing a

virtualization project; managing thevirtualized infrastructure; systemadministration

Altiris SVS, 18Amazon Web Services (AWS), 92–93AMD

AMD-V (virtualization-enabled) chip, 28, 59, 118, 166, 183–184

multicore processors, 46appliances, 87, 93–94. See also virtual

appliancesapplication packaging, 17–18Application Service Providers (ASPs), 91application streaming, 18–19applications of virtualization (overview).

See also specific applicationsdevelopment and testing, 62–65disaster recovery, 77–78failover/high availability/load balancing,

69–76, 130server consolidation, 67–69, 130server pooling, 76–77, 130suitability for shared storage, 234–236table summarizing products, 79–81training, 65–67

applications, software. See softwareapplications

approval. See also business case forvirtualization; cost-benefit analysis

justifying hardware costs, 147–148obtaining for virtualization architecture,

141–142

AppStream’s Virtual Image Distribution, 19architecture creation

charting the proposed architecture, 141, 142

end result of, 139hardware requirements for redundancy,

141importance of, 140obtaining approval, 141–142overview, 134reviewing to define infrastructure design,

324, 332use cases review for, 140virtual server organization, 140–141

Art.com site, 39–41ASPs (Application Service Providers), 91authorization. See approval; business case

for virtualization; cost-benefit analysisautomated migration, 194–196, 203–207, 327AWS (Amazon Web Services), 92–93

• B •back-end drivers, 58backup

challenges with DAS, 29, 43infrastructure setup before migration, 197before migration, 193during pilot implementation, 150with server consolidation, 68simplified with storage virtualization, 44soft costs in the data center, 116

bare-metal approach, 21, 55–56, 210, 291Bellard, Fabrice (QEMU creator), 312binary translation, 24, 54–55BitTorrent For Dummies (Gardner and

Krug), 262BitTorrent peer-to-peer file sharing, 262, 268blogs about virtualization, 337Boeing Dreamliner project, 14

Index

27_148310 bindex.qxp 11/1/07 12:00 AM Page 341

COPYRIG

HTED M

ATERIAL

business case for virtualization. See alsocost-benefit analysis

energy costs lowered, 44–45hardware costs lowered, 34–37importance of, 332IT operational flexibility increased, 37–41IT operations costs reduced, 41–44problems solved by virtualization, 33–34software licensing cost challenges, 45–48

business problems, identifying, 136business requirements, identifying,

136, 322–323

• C •CAD (Computer-Aided Design), 14Capacity Planner (VMware), 147capacity planning

defined, 146historical importance of, 164key server resources, 164–175for memory, 172–174for NICs, 167–168overview, 146–147performance issues, 165for storage, 168–172tools for, 147virtualization and need for, 146–147,

164–165CA’s Unicenter software manager, 195celebrating your success, 334central processing unit (CPU). See

processorscentralized storage. See storage

virtualizationcharting the proposed architecture,

141, 142chips. See processors; x86 chipsCitrix’s Presentation Server, 19client virtualization

application packaging, 17–18application streaming, 18–19defined, 17hardware emulation, 19–20reasons for pursuing, 17

cloningavoiding for migration, 201system, shared storage for, 234

clustering. See also HA (high availability)defined, 71, 130drawbacks of, 74DRS versus, 77products providing, 75

clustering software for storage, 230commodity hardware, 19, 50, 159Computer-Aided Design (CAD), 14concurrent users, licensing practices

related to, 47containers. See operating system

virtualization (containers)Converter migration tool (VMware),

152, 204, 205cost-benefit analysis. See also hard costs;

soft costscreating your spreadsheet, 122–127defining your virtualization solution, 112establishing current costs, 112–116hard cost reductions, 120–122identifying benefits of virtualization,

103, 119–122identifying new hardware needs, 117–118identifying virtualization project costs,

116–119net present value (NPV), 123project approval speeded by, 117, 127selecting a deployment scenario, 116–117for shared storage, 241soft cost reductions, 122steps for performing, 111–112usefulness of, 111, 117, 127

costs. See also cost-benefit analysis;energy costs; licensing

evaluating for virtualization software, 145investment in virtualization, 104–105, 110justifying for hardware, 147–148NAS cost-effectiveness, 30operations, reduced by virtualization,

41–44of overtime, 41–42pilot implementation for saving, 148

342 Virtualization For Dummies


of poor planning, 134–135price elasticity, 212of SANs, 31server consolidation savings, 36–37, 43of shared storage, 241“sticker shock” during implementation,

133of storage implementation, 231, 241system administration, 15–16, 34, 101, 122as virtualization drivers, 14–16, 34, 101of virtualization management, 216, 219

CPU (central processing unit). Seeprocessors

crashes. See operating system crashes;outage recovery

CRM (Customer RelationshipManagement), 53, 90, 264–266

• D •DAS (Direct-Attached Storage)

administrative complexity with, 233data management drawback with, 43defined, 29flexibility issues, 233increasing capacities with, 224–226limit of total data per server, 43, 232–233limitations for virtualization, 43, 169–170,

198, 228, 232–233as local storage, 224NAS and SAN versus, 169–170overview, 224–226performance issues, 169scalability issues, 228, 233virtualization applications using, 232VMware products using, 209wasted data capacity with, 228

data centershard costs in, 114–115hardware emulation use by, 19managing virtualization as equal member,

213, 215–216, 221“one application, one server” approach,

37–38, 137, 160–163, 165“orphan” applications in, 35

remote, 91–93soft costs in, 115–116space issues as virtualization driver,

13–14, 20, 33database server virtualization, avoiding,

107–108, 331Debian Linux. See Linuxdeciding to use server virtualization. See

also cost-benefit analysisapplications not suitable for, 107–109benefits, 101checking prerequisite conditions, 103–106company characteristics important for,

105–106determining financial benefits, 103investment issues, 104–105, 110power company support, 108–109resistance by employees, 105situations not suitable, 106–110, 330–331training requirements, 104–105, 330vendor support issues, 104, 107, 333

DeepBurner image burner, 269departmental implementation, 217–218,

220–221development and testing environments

crashes in, 63, 64DAS in, 232hardware emulation for, 25multiple OS requirements, 62–63virtualization benefits for, 63–64virtualization products for, 65, 79

device driversfuture improvements, 186hardware emulation issues, 25–26, 56OS virtualization issues, 52paravirtualization advantages for, 27, 58stub drivers with paravirtualization, 58

device virtualization, 185Direct-Attached Storage. See DASDirector system manager (IBM), 204Distributed Resource Scheduler (DRS) by

VMware, 77DNS (Domain Name Server), 172–173,

256, 295Docs and Spreadsheets (Google), 53

343Index

27_148310 bindex.qxp 11/1/07 12:00 AM 43

DR (disaster recovery), 77–78, 81. See alsooutage recovery

Dreamliner project of Boeing, 14drivers, device. See device driversdrivers of virtualization

data center space issues, 13–14, 33energy costs, 14–15, 34, 101hardware underutilization, 11–13,

33–37, 101paravirtualized drivers, 312–314system administration costs, 15–16,

34, 101DRS (Distributed Resource Scheduler) by

VMware, 77

• E •EC2 (Elastic Compute Cloud), 92, 93efficiency. See performanceElectronics Magazine (April 1965), 12EMC, VMware acquired by, 1encapsulation, 10End User License Agreement. See EULA;

licensingend users

application streaming transparent to, 18licensing practices related to, 47problems caused by, 17proprietary attitude toward servers by,

35–36, 137–138software installation by, 17, 18virtual appliance benefits for, 88, 210

energy costsbusiness case for virtualization, 44–45in cost-benefit analysis, 114, 120–122difficulties tracking, 44hard costs in the data center, 114power company support for

virtualization, 108–109rebate programs for virtualization, 109reduced by new chips, 45reduced by virtualization, 44–45,

101, 120–122as virtualization driver, 14–15, 34

ERP (Enterprise Resource Planning)application (NetSuite), 53

ESX Server (VMware)bare-metal approach of, 210Converter support for, 205for development and testing

environments, 65external storage types for, 170–171, 175for failover/high

availability/clustering/load balancing, 75

hardware emulation provided by, 26, 56PXE for installing, 196redundant NICs for, 168, 175for server consolidation, 68–69for server pooling, 77for training environments, 66VMware Server versus, 246, 291

EULA (End User License Agreement). Seealso licensing

VMware Server, 248, 249Windows XP SP2, 304XenExpress OS, 294XenExpress PV drivers, 312

evaluating server virtualization. Seedeciding to use server virtualization

experimentation phase, 217, 220, 221,329–330

• F •failover. See also HA (high availability);

load balancing; SPOF (single point of failure)

components of, 71–72defined, 39, 70drawbacks of mirroring for, 71hypervisor role in, 69–70, 72importance of, 70–71issues affecting availability, 70for mission critical applications, 70products providing, 75, 80shared storage for, 235–236simple case, 72, 130vendor terms for, 70virtual machine state for, 71–72virtualization benefits for, 71



failures, hardware. See DR (disasterrecovery); failover; outage recovery;SPOF (single point of failure)

Fedora Free Media Program, 268Fedora virtualization

Fedora Linux on VMware Server, 258–263installing a guest OS, 287–290installing a guest virtual machine, 278–287installing Fedora 7, 269–278obtaining Fedora 7, 268–269paravirtualization versus full

virtualization, 282password for, 271–272Red Hat Enterprise Linux features with,

267, 291virtualization-enabled chips with, 282

Fibre Channelcables for, 227with ESX Server, 170–171HBAs (Host Bus Adapters) for, 170, 227iSCSI protocol for, 228overview, 227preparing for migration, 198protocols, 30, 31SCSI used for data communication, 227specialized protocol required for, 170–171

file structure, defined, 51Fire 4600 server (Sun), 180Fire X4200 server (Sun), 181flexibility

DAS issues, 233of hardware emulation, 25of infrastructure, virtualization for, 106need for, 37operational, increased by virtualization,

37–41as VMware Server advantage, 247

free virtualization products. See alsoVMware Server; Xen (XenSource)

managing, 212–213obtaining, 336OpenSolaris (Sun), 61, 65, 67, 68OpenVZ (SWsoft), 23, 61, 67overview, 61

future of virtualizationbetter device support, 186built-in to dominant OSes, 61

device virtualization, 185diffusion into technology ecosystem, 83diffusion into the Internet, 90–93for high availability, 75–76implementation options, 61integration into OSes, 84–85, 93–94,

210–211IT personnel skill set changes, 93–95management as distinction between

products, 84memory improvements, 184NIC improvements, 184, 186preinstallation on virtual appliance,

85–89, 93–94remote data centers, 91–93SaaS, 90–91server pooling as, 77for software pricing practices, 95–97storage evolution, 240

• G •Gardner, Susannah (BitTorrent For

Dummies), 262Google’s Docs and Spreadsheets, 53graphics-intensive applications, 108grid architectures, 109growth, virtualization as aid to, 105guest operating systems

applications contained in, 21defined, 10in development and testing

environments, 64in hardware emulation, 19–20, 24, 25hypervisor interaction with, 24installing Fedora image files on VMware

Server, 261–263installing on Fedora 7, 287–290installing on VMware Server, 258–261installing Windows XP SP2 on

XenExpress, 300–311isolated by VMMs, 25modification needed for

paravirtualization, 28, 57, 59privileged guest, 57

345Index


• H •HA (high availability). See also failover;

load balancingdefined, 130future directions, 75–76hypervisor role in, 69–70, 76issues affecting availability, 70overview, 73products providing, 75, 80shared storage for, 235–236vendor terms for, 70

hard costs. See also cost-benefit analysischargebacks, 113for current infrastructure, 112–115in data center, 114–115defined, 113documentation easier for, 112–113energy or power, 114, 120–122hardware and physical equipment,

117–118hardware maintenance, 114–115, 120identifying reductions with virtualization,

120–122keeping separate from soft costs, 113licensing, 118, 119, 120outside services, 115in spreadsheets for cost-benefit analysis,

123–124training, 119for virtualization project, 117–119

hard drive partitions, 228hardware. See also hardware emulation;

outage recovery; specific typesappliances, 87–89, 93–94application requirements and suitability

for virtualization, 108business case for virtualization, 34–37,

41–44capacity planning, 146–147, 164–165commodity hardware, 19, 50, 159cost justification, 147–148costs lowered by virtualization, 34–37,

41–44defined, 50

designed for virtualization, 45, 175,180–181, 182

headroom, 101identifying virtualization project costs,

117–118incompatibility issues, 216keeping up on developments, 337–338licensing practices related to, 46low-wattage chips, 45management choices affected by, 213“one application, one server” approach,

37–38, 137, 160–163, 165price/performance curve, 34–35redundancy, planning for, 141risk exposure with virtualization, 163–164selecting when planning virtualization,

132, 145–148, 325–326server resources, 164–175storage type impacting server choices,

171–172underutilization as virtualization driver,

11–13, 33–37, 101underutilization rate for servers, 35virtualization product choice affected by,

145, 174, 333–334virtualization-enabled chips, 28, 59, 118,

166, 183–184VMware Server pros and cons, 246–247

hardware emulation. See also migrating tovirtualization

bare-metal approach, 21, 55–56benefits of, 23, 25, 55binary translation by, 24, 54–55in client virtualization, 19–20data center use of, 19defined, 19device driver issues, 25–26, 56drawbacks of, 25–26, 56emulated hardware layer, 55flexibility of, 25hypervisor installation approaches, 55–56illustrated, 24migration between hypervisors with, 23OS virtualization versus, 21, 22, 51, 60overview, 19–20, 23–26, 54–56, 129



paravirtualization versus, 27, 60, 129performance penalty with, 25, 56products providing, 20, 26, 56, 60in server virtualization, 23–26storage options, 237–238uses for, 25

hardware failures. See DR (disasterrecovery); failover; outage recovery;SPOF (single point of failure)

HBAs (Host Bus Adapters)defined, 41described, 170as NAS interface device, 170preparing for migration, 198redundant, 230as SAN interface device, 30, 170, 227

high availability. See HAhorizontal scaling, shared storage for,

234–235host system, defined, 10HP

OpenView software manager, 195System Insight Manager, 204

hypervisors. See also hardware emulationbinary translation by, 24, 54–55defined, 23, 54emulated hardware layer for, 55future of, 85guest OS interaction with, 24installation approaches for, 55–56memory requirements, 173–174migration between, 23role in failover, 69–70, 72role in high availability, 73, 76security issues, 198–199tracking performance after migration, 208

• I •IBM

Director system manager, 204System x3550 server, 180Tivoli software manager, 195transaction count licensing method, 47

IDE (Integrated Drive Electronics), 225

implementing virtualizationcelebrating your success, 334confirming planning assumptions and

conclusions, 143expecting change, 331–332hardware selection, 132, 145–148, 325–326life cycle phase, 132–133managing the infrastructure, 133, 154–157migrating to production environment,

133, 152–153options, 61pilot implementation, 148–151, 326production environment implementation,

326software product selection, 132,

144–145, 325steps for, 143“sticker shock” during, 133

InfoWorld virtualization newsletter,336–337

Infrastructure as a Service, 92–93installation

approaches for hypervisors, 55–56customized management products for,

195–196of Fedora 7, 269–278of Fedora image files on VMware Server,

261–263future of virtualization, 85–89of guest OS on Fedora 7, 287–290of guest OS on VMware Server, 258–261of infrastructure for virtualization,

197–200paravirualized drivers, 312–314prebooting solutions, 196of software by end users, 17, 18software management products for,

195, 203tasks for system administrators, 86third-party tools for, 152–153, 203vendor support issues, 86–87vendor-specific tools for, 152, 203virtual appliances for easing, 87–89of virtual appliances on VMware Server,

263–266

347Index


installation (continued)of virtual machine on Fedora 7, 278–287of virtual machine on VMware Server,

253–258virtualization implementation options, 61of virtualization software on server,

automated, 194–196, 203–207, 327of virtualization software on server,

manual, 153, 194, 202–203, 327of VMware Server, 248–251

Integrated Drive Electronics (IDE), 225Intel

multicore processors, 46Preboot Execution Environment

(PXE), 196VT (virtualization-enabled) chip, 28, 59,

118, 166, 183–184Internet

SaaS (Software as a Service), 53, 90–91storage requirements increased by use

of, 13virtualization diffusion into, 90–93

Internet resourcesblogs about virtualization, 337DeepBurner image burner, 269Fedora 7 download, 268Fedora image files, 262Google’s Docs and Spreadsheets, 53InfoWorld virtualization newsletter,

336–337Invirtus, 204Leostream, 204PG&E rebate program, 109, 127PlateSpin, 147, 204P2V vendors, 204SearchServerVirtualization site, 336storage virtualization information, 339vendor information, 339virtualization information, 336–339virtualized pricing information, 97VMware Capacity Planner, 147VMware Server appliances, 89, 264VMware Server download, 248VMware user group listings, 338XenExpress download, 292

Intrepid management product (Levanta), 196

Invirtus migration tools, 204IP addresses

for remote data centers, 93for virtual containers, 52VMware Server options, 255–256for XenServer, 299

iSCSIwith ESX Server, 170, 171network traffic increased by, 31overview, 171, 228preparing for migration, 198SCSI used for data communication, 228standard corporate network used by, 30standard IP Protocols used for, 171subnetting for, 171

IT. See also system administrationfuture of, 1–2, 93–95higher-level management with

virtualization, 94–95job dissatisfaction in, 42operational flexibility increased by

virtualization, 37–41operations costs reduced by

virtualization, 41–44proprietary attitude toward servers by,

35–36, 137–138system administration costs as

virtualization driver, 15–16, 34, 101work reduced by application

streaming, 18

• J •JBOD (just a bunch of drives), 169, 229

• K •Krug, Kris (BitTorrent For Dummies), 262

• L •Leostream migration tools, 204Levanta’s Intrepid management

product, 196



licensingin cost-benefit analysis, 118, 119, 120for Fedora 7, 275future of, 95–97general practices, 46–47OS virtualization benefits, 22total licenses as bargaining point, 48virtualization challenges for, 45, 47–48,

95–97VMware Server EULA, 248, 249Web site for virtualized pricing

information, 97Windows XP SP2 EULA, 304XenExpress EULA, 294XenExpress PV drivers EULA, 312

life cycle. See virtualization life cycleLinux. See also Fedora virtualization; Xen

(XenSource)automated migration options, 205blogs about virtualization, 337file structure, 51installing Fedora image files on VMware

Server, 261–263installing Fedora OS on VMware Server,

258–261managing virtualization as specialized

resource, 214manual migration for, 202–203Microsoft Server virtualization with, 61paravirtualization with, 59QEMU technology for virtualization-

enabled chips, 311–312virtualization features, 61workgroups in organization structure,

138–139Xen inclusion in distributions, 28, 84

load balancing. See also failover; HA (highavailability); server pooling

defined, 39, 130hypervisor role in, 69–70overview, 75products providing, 75, 80redundant NICs for, 168as SPOF protection, 74storage virtualization required for, 75vendor terms for, 70

local storage. See DAS (Direct-AttachedStorage)

LUNs (Logical Unit Numbers), 229, 230

• M •Macintosh platform, hardware emulation

for, 20mainframes, virtualization origins on, 11maintenance and upgrades. See also

backup; system administrationbenefits of virtualization for, 42, 120in cost-benefit analysis, 114–115, 120hard costs in the data center, 114–115before migration, 193during normal business hours, 42during off-hours, 41–42patch propagation with OS

virtualization, 53storage choices impacting, 240

malware, client virtualization reducing, 17managing a virtualization project

architecture selection, 134, 139–142, 324, 332

celebrating your success, 334hardware selection, 145–148implementing your solution, 142–157managing the infrastructure, 154–157migrating to production environment,

152–153organizational structure evaluation, 134,

137–139, 323–324pilot implementation, 148–151, 326planning assumptions and

conclusions, 143software product selection, 144–145types of virtualization, 129use cases identification, 134–137, 322–323virtualization application profile

choices, 130virtualization life cycle, 130–133virtualization plan creation, 134–142

managing the virtualized infrastructurechallenges of, 211–212commercial software for, 154–155costs of, 216, 219

349Index


managing the virtualized infrastructure(continued)

creating your own solution, 156–157current approach’s impact on, 218–219,

220–221current hardware mix issues, 219–220in departmental implementation,

217–218, 220–221as equal member of data center, 213,

215–216, 221expecting change, 331–332in experimentation phase, 217, 220, 221with free virtualization, 212–213importance of, 154, 327incompatibility issues, 216life cycle phase, 133multiple tools for, 156open source software for, 157philosophies of, 213in pilot implementation, 217–218, 220–221planning for, 133postponing virtualization, 156in production environment, 218, 221recommendations, 220–221solutions for, 155–157as specialized resource, 213, 214–215,

220–221storage administration, 229–230table summarizing recommendations,

220–221toolset choices for, 136, 154–155virtualization journey steps, 213, 217–218,

220–221virtualization software tools for, 84, 155

manual migration, 153, 194, 202–203, 327memory

amount per virtual machine, 173in blade servers, 179for Fedora 7 virtual machine, 280future technology, 184as key server resource, 172–174, 175requirements for virtualization, 172–17432-bit server issues, 177for XenExpress virtual machine, 302

Microsoft. See also Microsoft Servervirtualization; Microsoft Windows

blog site on virtualization, 337

dispute with VMware, 48hardware emulation offered by, 20, 26pricing model for virtualization, 96SoftGrid, 18, 19Softricity acquired by, 19virtual appliances from, 89Virtual Server, 20, 26, 65, 66Windows Server, 19

Microsoft Server virtualizationchild partition as guest OS, 57memory requirements, 173paravirtualization approach with,

28, 57, 61pricing model for, 96root partition as privileged guest, 57Virtual Server superceded by, 26virtualization integration into OS, 84

Microsoft Windowsfile structure, 51installing XP SP2 as guest OS on

XenExpress, 300–311malware attacks on, 17paravirtualization for, 28workgroups in organization structure,

138–139migrating to virtualization. See also

managing a virtualization projectautomated software installation for,

194–196, 203–207backing up applications and data, 193challenges of, 31–32, 189–190, 200–201cleaning up the source system, 192–193cloning for, avoiding, 201custom management products for,

195–196database options, 153file-based data options, 153free tools for, 205hardware inconsistencies, 201installing infrastructure, 197–200with Linux systems, 202–203manual software installation for, 153, 194,

202–203, 327migrating physical servers (P2V), 191,

200–207



moving to production (cutover), 191,207–208

options for, 152–153phases overview, 190–191planning for, 133planning needed for, 152, 153preparing the virtualized environment,

190–200preparing virtualization servers, 193–196production environment, 133,

152–153, 326removing hardware-dependent

software, 192removing unused and/or obsolete

software, 192security issues, 198–199storage choices impacting, 240system management tools for, 195, 203testing the infrastructure, 200third-party tools for, 152–153, 203time constraints and pressures during,

152, 153updating current OS and applications, 193vendor-specific tools for, 152, 203in virtualization life cycle, 133

migration. See also migrating tovirtualization

between hypervisors, 23of pilot systems, 149between servers, 42software for applications, 38

mirrored systems, 71mission-critical applications, shared

storage for, 235Moore, Gordon (Intel cofounder), 12, 13Moore’s Law, 12–13, 35, 164multicore processors, 46, 48, 165–166multimachine architectures, 109multiplexing with paravirtualization, 26, 57multisystem configurations with VMware

Server, 247

• N •NAS (Network-Attached Storage). See also

storage virtualizationcost-effectiveness of, 30

DAS versus, 169–170defined, 224drawbacks of, 30Host Bus Adapter (HBA) for, 170identifying virtualization project

costs, 118migration between servers with, 42network traffic increased by, 30overview, 29–30, 226–227scalability and reliability of, 227–228as shared storage, 226–227types of implementations, 169–170

NAT (Network Address Translation), 256net present value (NPV), 123NetSuite’s Enterprise Resource Planning

(ERP) application, 53Network Time Protocol (NTP), 295networking

Fedora 7 options, 284infrastructure preparation for

smigration, 198iSCSI-based SAN increasing traffic

over, 31NAS increasing traffic over, 30operating system use of, 51shared storage on networks, 226trend like Moore’s Law for, 13VMware Server options, 255–256XenExpress setup for, 295

New Energy data center, 181NFS (Network File System), 226NICs (Network Interface Cards)

defining for Fedora 7, 270for ESX Server implementations, 168future technology, 184, 186as key server resources, 167–168,

174–175not used by SANs, 30performance considerations, 167preinstalled on servers, 167redundant, advantages of, 168,

174–175, 23032-bit server issues, 176

Novell Linux. See LinuxNPV (net present value), 123NTP (Network Time Protocol), 295

351Index


• O •“one application, one server” approach

as dominant data center reality, 38for isolation of applications, 162limitations of, 37–38processors underutilized with, 165reasons for historical practice, 160–163resource contention avoided by, 161simplicity of, 160territoriality with, 137, 161–162

open source software. See also specificproducts

for development and testingenvironments, 65

free virtualization products, 61for server consolidation, 68for training environments, 67for virtualization management, 157

OpenSolaris (Sun), 61, 65, 67, 68OpenView software manager (HP), 195OpenVZ (SWsoft), 23, 61, 67operating system crashes. See also outage

recoveryin development and testing

environments, 63, 64OS virtualization dangers for, 52

operating system virtualization(containers). See also guest operatingsystems; migrating to virtualization

applications contained in virtual OS, 21benefits of, 22, 52density of containers with, 52device driver issues, 52ease of installation, 194environments suitable for, 23, 53–54free products, 61hardware emulation versus, 21, 22, 51, 60illustrated, 22IP addresses for containers, 52licensing requirements less with, 22limitations of, 23, 52–53multisystem configurations with VMware

Server, 247OS choice limited with, 23, 52–53overview, 21–23, 51–54, 129

paravirtualization versus, 51, 60patches propagated to all containers, 53performance benefits, 22, 52products providing, 23, 53, 60, 61storage options, 236–237terminology, 51–52uses for, 21–22, 53–54virtual OSes with, 21VMware products, 209

operating systems. See also guestoperating systems

choices limited with OS virtualization, 23, 52–53

described, 50integration of virtualization with, 84–85,

93–94, 210–211virtualization functionality with, 61

Oracle, licensing approach of, 47–48organizational structure

evaluating skill sets and experiencelevels, 139

need for understanding, 137, 332–333“one application, one server” approach,

137, 161–162reviewing before technical decisions,

323–324territoriality (hands off “my” server

attitude), 137–138Windows versus Linux workgroups,

138–139“orphan” applications in data centers, 35outage recovery. See also DR (disaster

recovery); SPOF (single point offailure)

failover for, 39, 69–72hardware requirements for redundancy,

141hardware risk exposure with

virtualization, 163–164high availability for, 73hypervisor role in, 69–70, 72load balancing for, 74products providing, 75server pooling, 39usual steps for, 38–39with virtualization, 39

overtime costs, 41–42



• P •Pacific Gas & Electric (PG&E), 15,

108–109, 127Parallels (SWsoft), 20paravirtualization. See also migrating to

virtualizationbenefits of, 27, 58described, 10, 21, 26–27, 129drawback of, 28, 59with Fedora 7, 282free products, 61full virtualization versus, 282, 311–312guest OS modification needed for,

28, 57, 59hardware emulation versus, 27, 60, 129illustrated, 27multiplexing by, 26, 57native device drivers with, 27, 58OS virtualization versus, 51, 60overview, 26–28, 57–59“para” in, 27performance benefits, 27, 52, 58privileged guest with, 57products providing, 28, 57, 60, 61shared memory used by, 27, 58storage options, 238–239stub drivers with, 58virtualization-enabled chips for, 28with XenExpress, 311–314

patches. See maintenance and upgradesperformance

application requirements and suitabilityfor virtualization, 107–108

capacity planning issues, 165DAS issues, 169hardware emulation drawbacks, 25, 56memory’s impact on, 172–174, 175NIC considerations, 167OS virtualization benefits, 22, 52paravirtualization benefits, 27, 52, 58of pilot implementation, assessing,

150–151QEMU penalty, 312VMware Server issues, 246, 247

PG&E (Pacific Gas & Electric), 15, 108–109, 127

physical to virtual. See migrating tovirtualization; P2V

pilot implementationagile systems development versus, 148assessing performance, 150–151bounded timeframe for, 150costs saved by, 148DAS in, 232debriefing session after, 151defined, 148gaining management support for, 148importance of, 326installing the mini-production

environment, 149managing, 149–150, 217–218, 220–221migrating systems, 149moving into production, avoiding, 150performing, 149–150selecting subset of production

environment for, 149trying more than one virtualization

product, 145planning a virtualization project

architecture selection, 134, 139–142, 324, 332

confirming assumptions and conclusions,143

costs of poor planning, 134–135expecting change, 331–332high-level tasks, 132, 134importance of, 321–322life cycle phase, 131–132lobbying and politics, 132organizational structure evaluation, 132,

134, 137–139, 323–324reviewing the plan, 143, 332use cases identification, 132, 134–137,

322–323PlateSpin

migration-analysis tool, 152–153, 204PowerRecon planning tool, 147Web site, 204

power costs. See energy costsPowerRecon planning tool (PlateSpin), 147

353Index


Preboot Execution Environment (PXE), 196Presentation Server (Citrix), 19price elasticity, 212process tables, defined, 51processors. See also x86 chips

assigning to XenExpress virtual machine, 302

in blade servers, 179commodity hardware with x86 chips,

19, 50, 159hardware selection issues, 166as key server resources, 165–166, 174licensing practices related to, 46, 47–48Moore’s Law for, 12–13, 35, 164multicore, 46, 48, 165–166multiple, for avoiding SPOF, 166, 174price/performance curve, 3564-bit servers, 175, 178–180, 181software emulation of x86 chips, 2432-bit servers, 175, 176–178, 181underutilization as virtualization driver,

12–13, 33–37underutilized in “one application, one

server” approach, 165virtualization-enabled, 28, 59, 118, 166,

183–184production environment

implementing, 326migrating to virtualization, 133, 152–153selecting subset for pilot implementation,

149virtualization management in, 218, 221

products. See virtualization softwareproducts

project management. See managing avirtualization project

proprietary attitude toward servers, 35–36,137–138

prototyping, DAS for, 232P2V (physical to virtual). See also

migrating to virtualizationautomated software installation for,

194–196, 203–207, 327challenges of, 200–201cloning for, avoiding, 201

custom management products for,195–196

database options, 153defined, 190file-based data options, 153hardware inconsistencies, 201with Linux systems, 202–203manual software installation for, 153, 194,

202–203, 327options for, 152–153planning needed for, 152, 153source system, 192system management tools for, 195, 203target system, 192third-party tools for, 152–153, 203time constraints and pressures during,

152, 153vendor-specific tools for, 152, 203

P2V wizard (XenSource), 204PXE (Preboot Execution Environment), 196

• Q •QEMU technology for virtualization-

enabled chips, 311–312quality assurance, hardware emulation

for, 25

• R •RAID (redundant array of inexpensive

disks), 229, 230RAM. See memoryRDP (Remote Desktop Protocol), 314–317Red Hat Enterprise Linux (RHEL), 267, 291Red Hat Fedora Linux. See Fedora

virtualization; Linuxredundancy. See also capacity planning

for avoiding SPOF risk, 166, 168, 174hardware requirements for, 141multiple processors, 166, 174of NICs, 168, 174–175, 230with RAID, 229, 230in storage infrastructure, 230



redundant array of inexpensive disks(RAID), 229, 230

Remember icon, 5Remote Desktop Protocol (RDP), 314–317resource contention, 161, 246RHEL (Red Hat Enterprise Linux), 267, 291root file system, defined, 51

• S •SaaS (Software as a Service), 53, 90–91,

92–93Salesforce.com’s CRM application, 53, 90SAN (Storage Area Network). See also

storage virtualizationcosts of, 31DAS versus, 169–170defined, 224Fibre Channel versus iSCSI, 31Host Bus Adapter (HBA) for, 30, 170, 227identifying virtualization project

costs, 118migration between servers with, 42network protocols for, 30overview, 30–31, 227–228as shared storage, 227–228switch fabric, 227types of implementations, 169–170

sandboxing, 18SATA (Serial Advanced Technology

Attachment), 225scalability, 175, 227–228, 233SCSI (Small Computer System Interface),

225, 227, 228, 229SearchServerVirtualization site, 336seasonal demand variations, 37, 39, 106security, 198–199, 275SELinux, 275Serial Advanced Technology Attachment

(SATA), 225server consolidation

backup process with, 68consolidation defined, 67costs reduced by, 36–37, 43defined, 31, 67, 130

as first step in virtualization, 31, 36, 39, 67hardware emulation for, 25hardware risk exposure with, 163–164overview, 67–69products providing, 68–69, 79–80proprietary attitude toward servers,

35–36ratio of retired to continuing servers, 36real-world example, 36–37typical project tasks, 67–68underutilization rate for servers, 35

server pooling. See also load balancingArt.com example, 39–41defined, 39, 130integration into OSes, 84outage recovery with, 39overview, 76–77products providing, 77, 80–81seasonal demand variations eased by, 39shared storage for, 236SPOF avoided by, 164storage virtualization required for, 76in times of company restructuring, 39

server virtualization. See also hardwareemulation; migrating to virtualization;operating system virtualization(containers); paravirtualization

applications not suitable for, 107–109benefits, 101checking prerequisite conditions, 103–106company characteristics important for,

105–106deciding about using, 101–110described, 10determining financial benefits, 103installing software manually, 153, 194,

202–203, 327installing software using automation,

194–196, 203–207, 327investment in virtualization, 104–105, 110as main virtualization arena, 20, 31organizing physical servers, 140–141overview, 20–28power company support for, 108–109resistance by employees, 105

355Index


server virtualization (continued)situations not suitable for, 106–110,

330–331storage type impacting hardware choices,

171–172table comparing types of, 60training requirements, 104–105, 330types of, 21–28, 49, 59–60vendor support issues, 104, 107, 333x86-based, 50–51

servers. See also server virtualization;specific products and applications

backing up before migration, 193blade servers, 179choices impacted by storage, 171–172choosing, 175–182designed for virtualization, 45, 175,

180–181, 182existing, cleaning up before migration,

192–193key resources, 164–175migration between, benefits of

virtualization for, 42new, preparing for migration, 193–196NICs preinstalled on, 167“one application, one server” approach,

37–38, 137, 160–163, 165proprietary attitude toward, 35–36,

137–138reusing existing 32-bit hardware, 175,

176–178, 18164-bit, 175, 178–180, 181table summarizing choices, 181–182total data per server limitation with

DAS, 43underutilization rate for, 35

shared memory, 27, 58shared storage. See NAS (Network-

Attached Storage); SAN (Storage AreaNetwork); storage virtualization

Simple Storage Service (S3), 92single point of failure. See SPOF64-bit servers, 175, 178–180, 181–182SLES (SUSE Linux Enterprise Server), 214.

See also Linuxsmall business virtualization, 337

Small Computer System Interface (SCSI),225, 227, 228, 229

soft costs. See also cost-benefit analysisbackup, 116for current infrastructure, 112–114,

115–116in data center, 115–116defined, 113difficulties establishing, 113estimating, 122identifying reductions with

virtualization, 122keeping separate from hard costs, 113preponderance of, 112, 113–114in spreadsheets for cost-benefit analysis,

123–124system administration, 115–116, 122

Softgrid for Desktops (Softricity), 19SoftGrid (Microsoft), 18, 19Softricity’s Softgrid for Desktops, 19software applications

backing up before migration, 193contained in virtual OS, 21data generation by, 225end user installation issues, 17, 18evaluating support for, 144installation headaches, 86–87migration software for, 38“one application, one server” approach,

37–38, 137, 160–163, 165“orphans” in data centers, 35SaaS environments, 53suitability for virtualization, 107–109user-level, defined, 50

Software as a Service (SaaS), 53, 90–91,92–93

software development. See developmentand testing environments

software licensing. See licensingsoftware management products, 195, 203software products. See software

applications; virtualization softwareproducts

Solaris OS (Sun)for development and testing

environments, 65



for failover/highavailability/clustering/load balancing, 75

OS virtualization with, 23, 53for server consolidation, 68for training environments, 67

specialized resource, managingvirtualization as, 213, 214–215, 220–221

SPOF (single point of failure). See alsofailover

blade servers for avoiding, 179defined, 70failover for avoiding, 70, 74, 164load balancing as protection against, 74multiple processors for avoiding, 166, 174processor issues, 166RAID for avoiding, 229redundant NICs for avoiding, 168, 174risk exposure with virtualization, 163–164server pooling for avoiding, 39, 164

spreadsheets for cost-benefit analysisassumptions for sample project, 125benefits shown by, 125–127Current Cost Structure, 123, 124hard cost and soft cost sections, 123–124improving the project case, 126–127justifying financial projections, 126NPV (net present value) in, 123Project Cost Summary, 123, 125, 126reexamining assumptions, 127sample set, 123–125showing financial implications across

time, 123Total Yearly Hard Costs in, 126usefulness of, 122, 125Virtualized Cost Structure, 123, 124

S3 (Simple Storage Service), 92storage. See also specific types

administration, 229–230approaches to, 29–31assigning to XenExpress virtual

machine, 302backup problems, 29in blade servers, 179bottlenecks, 29costs of implementing, 231, 241

data redundancy, 230defined, 224hardware emulation options, 237–238importance for virtualization, 223increasing capacities for, 224–226JBOD (just a bunch of drives), 169, 229as key server resource, 168–172, 175LUNs (Logical Unit Numbers), 229, 230network, 226OS virtualization options, 236–237paravirtualization options, 238–239RAID (redundant array of inexpensive

disks), 229, 230requirements for virtualization, 168–172server hardware choices impacted by,

171–172shared, types of, 226–228space issues as virtualization driver,

13–14, 28, 3332-bit server issues, 176trend like Moore’s Law for, 13

Storage Area Network. See SANstorage virtualization. See also storage;

specific types of storageapplications suitable for shared storage,

234–236approaches to storage, 29–31clustering software, 230cost effectiveness of, 241DAS limitations for, 43, 169–170,

198, 228, 232–233data backup simplified with, 44external storage protocols, 170–171for failover, 235–236with hardware emulation, 237–238for high availability, 235–236horizontal scaling, 234–235infrastructure setup before migration, 198integration into OSes, 84IT costs lowered by, 43–44for large numbers of VMs, 235for load balancing, 74migration between servers with, 42need for, 28–29with OS virtualization, 236–237overview, 28–31

357Index


storage virtualization (continued)with paravirtualization, 238–239ramifications for networking and system

administration, 43for server pooling, 76, 236shared storage for, 234–236system cloning, 234types of implementations, 169–170virtualization journey steps, 239–240Web site for, 339

stub drivers, 58SugarCRM installation on VMware Server,

264–266Sun. See also Solaris OS (Sun)

Fire 4600 server, 180Fire X4200 server, 181OpenSolaris, 61, 65, 67, 68OS virtualization offered by, 23, 53

support issuesevaluating when choosing products,

144, 333for software installation, 86–87for virtualized software, 104, 107

SUSE Linux Enterprise Server (SLES), 214.See also Linux

SVS (Altiris), 18switch fabric for SANs, 227SWsoft. See also Virtuozzo (SWsoft)

OpenVZ, 23, 61, 67OS virtualization offered by, 23, 53Parallels, 20VZP2V migration tool, 204

Symantec’s Veritas Provisioning Manager, 196

system administration. See also IT;maintenance and upgrades

in cost-benefit analysis, 115–116, 122costs as virtualization driver, 15–16,

34, 101soft costs in the data center, 115–116

system cloning, shared storage for, 234System Insight Manager (HP), 204system libraries, defined, 52System x3550 server (IBM), 180

• T •Technical Stuff icon, 5Technology Without An Interesting Name

(TWAIN), 229territoriality

proprietary attitude toward servers,35–36, 137–138

Windows versus Linux workgroups,138–139

testing applications. See development andtesting environments

thin client information, 337Thinstall’s Virtualization Suite, 1832-bit servers, 175, 176–178, 181–182Tip icon, 5Tivoli software manager (IBM), 195torrents, 262training environments

as common virtualization application,65–66

OS virtualization suitability for, 53–54typical scenario, 66virtualization products for, 66–67

training requirements for virtualizationavoiding skimping on, 330in deciding about virtualization, 104–105identifying costs of, 119

transaction count licensing method, 47TWAIN (Technology Without An Interesting

Name), 229

• U •Ubuntu Linux as guest OS, 288–290Unicenter software manager (CA), 195Unix file structure, 51upgrades. See maintenance and upgradesuse cases

for architecture selection, 140asking people for, 137business problems addressed by, 135defined, 323for desired management toolset, 136



for desired virtual machine density, 136identifying, defined, 134importance of, 134–135, 322–323, 332items to evaluate when identifying,

135–136need for, 132for overall business requirements, 136for potential future requirements, 136

user-level applications. See softwareapplications

• V •vendor support issues. See support issuesVeritas Provisioning Manager

(Symantec), 196virtual appliances

benefits for end users, 88, 210benefits for vendors, 88, 210defined, 87as future of virtualization, 89impact on IT personnel, 93–94installing on VMware Server, 263–266Microsoft offerings, 89need for, 87VMware offerings, 89

Virtual Image Distribution (AppStream), 19Virtual Infrastructure 3 (VI3) by VMware

Art.com server pooling using, 40bare-metal approach of, 210DR capability of, 78DRS (Distributed Resource Scheduler), 77for failover/high


Virtual Iron, 69, 75, 77, 78Virtual Machine Manager (Red Hat), 214virtual machine monitor (VMM), 23, 25virtual machines. See VMsvirtual OSes. See guest operating systems;

operating system virtualization(containers)

Virtual PC Guy blog site, 337Virtual Server (Microsoft), 20, 26, 65, 66virtualization events, 338–339

virtualization life cycledefined, 32graphical representation, 131infrastructure, 327implementation phase, 132–133management choices affected by, 213,

217–218, 220–221need for understanding, 130–131operating phase, 133planning phase, 131–132storage methods in, 239–240

virtualization pilot. See pilotimplementation

virtualization software products. See alsofree virtualization products; specificproducts and companies

for application packaging, 18for application streaming, 19application-level, 291for automated software installation,

195–196, 204–207custom management products, 195–196for development and testing

environments, 65, 79for disaster recovery, 78, 81evaluating application software

support, 144evaluating costs, 145for failover/high

availability/clustering/load balancing,75, 80

hardware choices affecting, 145, 174, 181–182, 333–334

for hardware emulation, 20, 26, 56, 60management choices affected by, 213management systems with, 84, 155for OS virtualization, 23, 53, 60, 61for paravirtualization, 28, 57, 60, 61, 65, 67pilot program for testing, 145selecting for virtualization

implementation, 132, 144–145, 325for server consolidation, 68–69, 79–80for server pooling, 77, 80–81software management products, 195, 203,

216, 219table summarizing, 79–81for training environments, 66–67

359Index


Virtualization Suite (Thinstall), 18virtualization technology types, 60virtualization-enabled chips

Fedora virtualization with, 282full virtualization with, 282, 311–312hardware selection issues, 166Linux QEMU technology for, 311–312need for, 118overview, 28, 59, 183–184for Xen-based environment, 166with XenExpress, 311–312

virtualized storage. See storagevirtualization

Virtuozzo (SWsoft)for failover/high


OS virtualization with, 23, 53, 61for server consolidation, 68for training environments, 67

VI3. See Virtual Infrastructure 3 (VI3) byVMware

VMM (virtual machine monitor), 23, 25VMs (virtual machines)

accessing Windows guest VM with RDPclient in XenExpress, 314–317

determining desired density, 136in development and testing

environments, 64in hardware emulation for client

virtualization, 20in hardware emulation for server

virtualization, 23for high availability, 73installing on Fedora 7, 278–287installing on VMware Server, 253–258installing on XenExpress, 300–311IT management of, 94managing during pilot implementation, 150memory requirements, 173mixed workload issues, 138–139price elasticity of, 212security issues, 199shared storage for large numbers, 235state of, 71–72

VMware. See also ESX Server (VMware);Virtual Infrastructure 3 (VI3) byVMware; VMware Server

bare-metal products, 210blog site on virtualization, 337Capacity Planner, 147challenges from virtualization integration

into OSes, 84–85, 210–211Converter migration tool, 152, 204, 205dispute with Microsoft, 48DRS (Distributed Resource Scheduler), 77EMC acquisition of, 1hardware compatibility list, 118hardware emulation offered by, 26, 56IPO of, 1as leader in virtualization deployment, 61managing virtualization as specialized

resource, 214OS virtualization products, 209user group listings, 338virtual appliance offerings, 89VMware Workstation, 209

VMware Serveracquiring your free product, 248for development and testing

environments, 65ESX Server versus, 246, 291flexibility of host OS with, 247hardware emulation provided by,

20, 26, 56installing a guest OS, 258–261installing a virtual machine, 253–258installing Fedora image files, 261–263installing virtual appliances, 263–266installing VMware Server, 248–251memory requirements, 173multisystem configuration capabilities,

247performance issues, 246, 247pros and cons of, 246–247redundant NICs for, 175resource contention with, 246serial number for, 248starting the Server Console, 252–253for training environments, 66



virtual appliances, 89, 264virtualization approach of, 209, 246Xen versus, 291

VMware Workstation, 209VMworld event, 339VZP2V migration tool (SWsoft), 204

• W •Warning! icon, 5Web resources. See Internet resourcesWeb-hosting companies, OS virtualization

for, 21–22Windows. See Microsoft WindowsWindows Server (Microsoft), SoftGrid to be

incorporated with, 19

• X •x86 chips. See also processors

blogs about virtualization, 337commodity hardware based on, 19, 50, 159limitations of, 50, 54software emulation of, 24virtualization architected for, 50–51x86-based systems, 50

Xen (XenSource). See also Fedoravirtualization

AWS’s use of, 92bare-metal approach of, 291blog site on virtualization, 337described, 291for development and testing

environments, 65Domain0 as privileged guest, 57DomainU as guest OS, 57for failover/high


as free virtualization product, 61functionality extensions, 61hardware requirements, 117–118inclusion in Linux distributions, 28, 84managing virtualization as specialized

resource, 214

memory requirements, 173as paravirtualization example, 28, 57PXE for installing, 196for server consolidation, 69for training environments, 66virtualization-enabled chips for, 166, 183VMware Server versus, 291

XenEnterprise (XenSource), 69, 75, 205XenExpress (XenSource)

accessing a Windows guest VM with anRDP client, 314–317

dual-boot mode not possible for, 294installing paravirtualized (PV) drivers,

311–314installing Windows XP SP2 as guest

virtual machine, 300–311installing XenConsole, 296–297installing XenExpress, 293–295obtaining, 292–293paravirtualization versus full

virtualization, 311–312for server consolidation, 69working with XenConsole, 297–299

XenServer (XenSource), 69, 75XenSource. See also specific products

described, 292paravirtualization products, 28, 57,

61, 65, 67P2V wizard, 204server consolidation products, 69

• Z •zealotry, avoiding, 117

361Index


Numerics • COPYRIGHTED MATERIAL · overview, 146–147 performance issues, 165 for storage,...

Documents

Transcript of Numerics • COPYRIGHTED MATERIAL · overview, 146–147 performance issues, 165 for storage,...