Windows Server Platform:Overview and Roadmap

Sean McGraneProgram ManagerWindows Server Platform Architecture Groupsmcgrane @ microsoft.comMicrosoft Corporation

Session OutlineSession Outline

Server Hardware TrendsTechnology and Processor Trends

Form Factors: Blade Servers

Windows Longhorn Server Direction

ReliabilityHardware error handling

Hardware partitioning

Application ConsolidationVirtualization

Call to Action

Resources

Server Technology TrendsServer Technology Trends

ProcessorsMore processing units per physical packageIn this presentation 1P means one physical processorPoint-to-point bus architectureDirect attached memory

Memory – capacity continues to increaseMemory technology is a feature of the processorFully Buffered DIMM (FBD) by 2007/2008

I/O – moves to PCI ExpressIncreased IO bandwidth and reliability

FirmwareIncreased adoption of Extensible Firmware Interface (EFI)

PlatformsIncreased adoption of blades for 1P/2P application loadsScale up moves to the commodity spaceLarge number of processing units on high-end servers (256 or more)

Processor TrendsProcessor Trends

WindowsWindowsLonghorn ServerLonghorn Server

CoreCore CoreCore CoreCoreCoreCore

CacheCache

Server 2003 SP1Server 2003 SP1

Pipeline/Pipeline/CachesCaches

ASAS ASAS

Per

form

anc

Per

form

anc

ee

Dual Dual ThreadThread

Dual Dual CoreCore

Quad Quad CoreCore

Higher number of Higher number of cores per processorcores per processor

All new server All new server processors are 64-processors are 64-

bit capablebit capable

Server 2003 Compute Server 2003 Compute Cluster EditionCluster Edition

TimeTime

Performance Performance capability of today’s capability of today’s

x86 8P serverx86 8P serverCacheCache

CoreCore CoreCore

What will customers do with Multi-Core?What will customers do with Multi-Core?

Typical application scaling can’t keep up1P and 2P servers are often under utilized today

Future 1P servers will be more compute capable than today’s 8P

Few customer loads fully utilize an 8P server today

Application consolidation will be the volume solutionMultiple application loads deployed to each processor

Scale up apps can be accommodated on volume servers

How will form factors be affected?IO & memory capability must match compute capability

IO expansion isn't available in today's volume server form factors

Larger form factors may be required for these servers

Can RAS scale with performance?Consolidation and scale-up apps raise RAS requirements

Mid- to high-end RAS features are needed on volume servers

Typical Blade Platform TodayTypical Blade Platform Today

Current models are typically 6U to 7U chassis with 10 to 14 1P/2P x64 bladesEach blade is like a server motherboard

IDE/SCSI attached disks, network and IO Daughter card on the bladeMidplane is passive; routing is very complex; IO switches provided in the chassisSAN attached rate is high, ~40%

Initial problems with adoptionCosts were too highLimited vendor network switches availableData center infrastructure not ready, cabling, management, power, etc

Aggregated server management potential not achievedProprietary interfaces to the management moduleStatic blade configurationOS state on the blade complicates repurposing

Compute BladesCompute Blades Chassis midplaneChassis midplane

Network switchesNetwork switches

FC switchesFC switches

Chassis Chassis Management Management

Module (CMM)Module (CMM)

1GBE NIC

1GBE NIC

CPU

CPU Memory

Chipset

IDE Drive

IDE Drive

Fiber Channel Daughter Card

1GBE NIC

1GBE NIC

CPU

CPU Memory

Chipset

IDE Drive

IDE Drive

Fiber Channel Daughter Card

1GBE NIC

1GBE NIC

CPU

CPU Memory

Chipset

IDE Drive

IDE Drive

Fiber Channel Daughter Card

1GBE NIC

1GBE NIC

CPU

CPU Memory

Chipset

IDE Drive

IDE Drive

Fibre Channel Daughter Card

Future Blade PlatformFuture Blade Platform

Similar chassis configuration, e.g. 6U to 7U chassis with 10 to 14 1P/2P x64 bladesThe compute blade becomes stateless

All IO and direct attached disks are removedConsolidated storage on FC or iSCSI SANMore reliable storage solution, reduces cost and simplifies managementSimplifies blade failover and repurposing

The chassis contains a set of configurable componentsThe midplane is PCIe only and contains a programmable PCIe switchAll IO devices and switches are at the far end of the midplaneThe CMM programs the PCIe switch to assign IO to compute blades, i.e. configure servers

Aggregated server management potential is realizedStandardized management interfaces implemented in the CMMFlexible and dynamic configuration of blade serversSimplified server repurposing on error; Failed components can be configured out

Compute BladesCompute Blades Chassis midplaneChassis midplane

Network IO/switchesNetwork IO/switches

FC IO/switchesFC IO/switches

Chassis Chassis Management Management

Module (CMM)Module (CMM)

Switch

CPU

CPU

Chipset

Memory

PCIECPU

CPU

Chipset

Memory

PCIECPU

CPU

Chipset

Memory

PCIECPU

CPU

Chipset

Memory

PCIe

Blade Support - Remote BootBlade Support - Remote Boot

Microsoft supports remote boot with Server 2003Supported for both FC and iSCSI SANSAN boot requires a Host Bus Adapter (HBA)Windows install processes work with this configuration

iSCSI creates a new low end SAN marketSoftware initiated install and boot is complexA low-cost HBA is a simpler approach Enables faster time to market solutionProvides a solution for exiting OSs, e.g. Server 2003

SAN management is too complexMust be simplified to create a volume solutionSimple SAN program addresses this simplificationPackaged SAN solutions with a single point of managementInitial focus is simplifying SAN deploymentSAN boot simplification is a longer term goal

Power and CoolingPower and Cooling

Processor power ratings and server density continue to riseHigh-end processors will have 130W footprintBlade servers can populate up to 168 procs per rack

Existing data center infrastructure can’t copeAt 65-95W per sq foot, can supply about 6-7KW per rackA single fully loaded blade chassis can be rated at >5KW

Power management can helpProcessor p-states supported in Server 2003 SP1Balances power consumption to real time utilizationTransparent to the user and applicationsCan lower processor power consumption up to 30%

More is needed, new power initiatives are emergingMore efficient power supplies with monitoring capabilitySilicon advances to reduce processor power leakageTools to accurately rate server power

Power and cooling are a huge customer problemPower management alone can't solve the problemUpgrades to legacy data center infrastructure will be required

Longhorn Server Platform DirectionLonghorn Server Platform Direction

Move the industry to 64-bit (x64) WindowsCompatibility for 32-bit apps on x64

Broad coverage for 64-bit drivers

Enable Windows on Itanium for scale up solutions

Consolidate multiple applications per serverHomogeneous consolidation for file, print, web, email, etc

Virtualization for heterogeneous low to mid-scale application loads

Hardware partitions for heterogeneous scale up application loads

Improve Reliability, Availability, and ServiceabilityHardware error handling infrastructure

Enhanced error prediction and redundant hardware features

Continue progress on Windows performance

Improved support for Windows operation on an iSCSI or FC SAN

WindowsWindows Hardware Error Architecture (WHEA) Hardware Error Architecture (WHEA)

Motivation - Improve reliability of the serverConsolidation raises server RAS requirements

Server 2003 bugcheck analysis:~10% are diagnosed as hardware errors

Others exhibit corruption that could be hardware related

Hardware errors are a substantial problem on serverSilent hardware errors are a big concern

OS participation in error handling is inconsistentImproved OS integration can raise server RAS level

GoalsProvide information for all hardware error events

Make the information available to management software

Reduce mean time to recovery for fatal errors

Enable preventative maintenance using health monitoring

Reduce crashes using error prediction and recovery

Utilize standards based hardware, e.g. PCIe AER

WHEA – The ProblemWHEA – The Problem

Lack of coordinated hardware error handlingDisparate error sources with distinct mechanismsError signaling and processing is architecture specificPoor I/O error handling capability; improved with PCIe AERLack of OS integration lowers server RAS

Lack of a common data format restricts OS participationNo mechanism to discover error sourcesSome hardware errors are not reported to the OSNo way to effectively utilize platform-specific capabilities

WHEA is a common hardware error handling infrastructure for Windows

Error source identification, configuration and managementCommon hardware error flow in WindowsPlatform driver model to provide hardware/firmware abstractionCommon hardware error record format for all platformsStandard interface to persist error recordsHardware error events provided to management software

Dynamic Hardware Partitioning (DHP)Dynamic Hardware Partitioning (DHP)

Memory

Memory Memory

Memory

IO Bridge

Service Processor

1. Partition Manager provides the UI for partition creation and

management

2. Service Processor controls the inter processor and IO

connections

Partition Manager

3. Hardware partitioning to the socket level. Virtualization for sub

socket partitioning

4. Support for dynamic hardware addition and replacement in

Longhorn Server

PCI Express

Core Core

Cache

… Core Core

Cache

…

Core Core

Cache

…Core Core

Cache

…

. . .

IO Bridge

. . .IO Bridge

. . .

IO Bridge

. . .

Longhorn Longhorn dynamic hardware dynamic hardware

partitioning partitioning features are features are focused on focused on

improving server improving server RASRAS

Future Hardware Partitionable ServerFuture Hardware Partitionable Server

DHP – Hot AdditionDHP – Hot Addition

Addition of hardware to a running partition with no downtimeProcessors, memory and IO subsystems may be added

Scenarios supported by Hot AdditionExpansion of server compute resourcesAddition of I/O extension unitsEnable unused capacity in the server

Hot Addition sequenceHardware is physically plugged into the serverAdministrator or management software initiates a Hot AdditionThe firmware initiates an ACPI Hot Add notify to the OS in the partitionThe OS reads the ACPI tables and utilizes the unit described by the notify

Operations are not transparent to applications or device driversA notification API will be made available for both user and kernel modeDrivers cannot assume hardware resources are static

Units are added permanentlyTo subsequently remove the unit requires a reboot of the partition

DHP – Hot ReplaceDHP – Hot Replace

A processor/memory unit is replaced with a redundant spareImplemented with no OS downtimeThe details of the Hot Replace sequence are being defined

System requirementsOne or more spare units in the serverHardware assistance can improve efficiency of the swap process

Scenarios supported with no downtimeReplacement of a unit initiated by hardware failure predictionReplacement of a unit by service engineers during maintenance

Hot Replace sequenceAdministrator or management software initiates a Hot ReplaceA spare unit is brought online and mapped into the partition viewFW initiates an ACPI replace notify to the OS which identifies the unitThe context of the unit to be replaced is migrated to the spare unitThe OS provides notification once the operation is completed Firmware maps out the replaced hardware without interruption to the OSThe OS completes the initialization of the new processors and continues

The operation is transparent to applications and device drivers

Microsoft View on PartitioningMicrosoft View on PartitioningUsed for server consolidation

Server consolidation: hosting multiple application loads on a single serverMicrosoft offers homogeneous consolidation programs for:

File, print, email, web, database, etcHeterogeneous side by side application execution is problematic

Applications tend to collide with each otherTesting is required to validate different application combinations

Partitioning offers out of the box server consolidation solutionsHardware Partitions

High levels of isolation and reliability with low perf overheadIdeal for scale up application consolidationGranularity of hardware is large; Removal of hardware is very complex

Software Partitions (Virtualization)Preferred direction for application consolidationFlexible partition configuration; granular dynamic Resource ManagementIdeal solution for consolidation of volume Windows applications

Future DirectionProvide a hypervisor based virtualization solutionExpand the application environments supported under virtualizationHardware partitions used for scale up application consolidation

Virtualization and Hardware PartitionsVirtualization and Hardware Partitions

Volume 32-bit application solutionOut of the box consolidationHeterogeneous OS/App consolidationSupported on standard serversHighly flexible and configurable solution64-bit Host support with VS 2005 SP1Host OS model not preferred forproduction deployment

Hardware partitioning provides physical isolationSoftware partitions may be used within a hardware partitionEnables software partitions and scale up application consolidation on a single server Requires partitionable hardware

Software Partitions using Virtual Server (VS) 2005

Hardware Partitions

Windows compliant server

Windows Host OS

App App App

NT4 Win2K Win2K3

Virtual Server

Windows Host OS

App App App

NT4 Win2K Win2K3

Virtual Server

Windows compliant partitionable server

App

Win2K3

Virtualization FuturesVirtualization FuturesOS virtualization layer replaced by a thinner hypervisor layer

Significant reduction in performance overhead and maintenanceMutli-processor support in the guest environment64-bit hypervisor to enable scaling

Devices can be assigned to a partitionRequires isolation protection support in the hardware (IO Virtualization)Partitions can share assigned device resource with other partitions

Higher levels of reliability and availabilitySnapshot of guest environment with no downtime enables high availability solutionsWHEA provides hardware health monitoring and higher levels of RASGuests can be moved between physical servers with no downtime

Granular and dynamic management of hardware resourcesManagement becomes a key differentiator in this environment

Enables heterogeneous high-availability and legacy production application consolidation on a non-hardware partitioned server

Windows compliant server

App App App

Win2K Win2K Win2K3

App

Win2K3

Hypervisor

App

Longhorn

App

Win2K3

Storage Network Storage Network StorageStorage

Call to ActionCall to Action

Server vendorsConsider the effect of multi core on volume serversConsider hardware partitions on mid range serversProvide management flexibility in blade chassisImplement power saving technologiesProvide WHEA extensions to improve server RASImplement dynamic hardware partitioning features to improve RASImplement emerging virtualization hardware assists

Device vendorsProvide 64-bit drivers for all devicesValidate compatibility in a dynamic hardware environment

ISVs – hardware management Implement to emerging standards based management interfacesProvide flexible blade chassis managementUtilize emerging power management standardsProvide enhanced RAS features based on WHEA information

Community ResourcesCommunity Resources

Windows Hardware & Driver Central (WHDC)www.microsoft.com/whdc/default.mspx

Technical Communitieswww.microsoft.com/communities/products/default.mspx

Non-Microsoft Community Siteswww.microsoft.com/communities/related/default.mspx

Microsoft Public Newsgroupswww.microsoft.com/communities/newsgroups

Technical Chats and Webcastswww.microsoft.com/communities/chats/default.mspx

www.microsoft.com/webcasts

Microsoft Blogswww.microsoft.com/communities/blogs

ResourcesResources

Blades and SANStorage track - Storage Platform leadership

Storage track – Simplifying SAN deployments on Windows

Networking track - Implementing convergent networking

Networking track - Network IO Architectures

http://www.microsoft.com/windowsserversystem/storage/simplesan.mspx

Reliability - Fundamentals track Windows Error Hardware Architecture (WHEA)

Error management solutions synergy with WHEA

Dynamic Hardware Partitioning

VirtualizationServer track – Virtual Server Overview and Roadmap

Fundamentals track – Windows Virtualization Architecture

Fundamentals track – Virtualization Technology for AMD Architecture

Fundamentals track – Virtualization Technology for Intel Architecture

http://www.microsoft.com/windowsserversystem/virtualserver/default.mspx