Windows Server 2016

Markus ErlacherCEOitnetX AGwww.itnetx.ch

Bringing the cloud to youWhat if you could have the control of the datacenter

and the power of the cloud?

Reduced developer chaos and shadow IT

Simplified, abstracted experience

Added consistency to modern development

One portal, one user experience

Write once, deploy anywhere

Nano Server

Voice of the Customer

- Reboots impact my business- Why do I have to reboot because of a patch to a component I never use?- When a reboot is required, the systems need to be back in service ASAP

- Server images are too big- Large images take a long time to install and configure- Transferring images consumes too much network bandwidth- Storing images requires too much disk space

- Infrastructure requires too many resources- If the OS consumes fewer resources, I can increase my VM density- Higher VM density lowers my costs and increases my efficiency &

margins

I want just the components

I needand nothing more

Microsoft Server Journey

Windows NT to Windows Server

2003

Windows/WindowsNT

Server Roles/Features

Windows Server 2008and

Windows Server 2008 R2

Server Core

Full Server

Windows Server 2012and

Windows Server 2012 R2

Server Core

Minimal Server Interface

GUI Shell

Windows NT / Windows Server 2003

Windows Server 2008 Windows Server 2008 R2

Windows Server 2012 Windows Server 2012 R2

Why we need Nano Server

- Azure- Patches and reboots interrupt service delivery- (*VERY large # of servers) * (large OS resource

consumption)- Provisioning large host images competes for

network resources

Why we need Nano Server

- Cloud Platform System (CPS)- Cloud-in-box running on 1-4 racks using System

Center & Windows Server- Setup time needs to be shortened- Patches and reboots result in service disruption

- Fully loaded CPS would live migrate > 16TB for every host OS patch

- Network capacity could have otherwise gone to business uses

- Reboots: Compute host ~2 minutes / Storage host ~5 minutes

We need server configuration optimized for the cloud

- A new headless, 64-bit only, deployment option for Windows Server

- Deep refactoring focused on - CloudOS infrastructure- Born-in-the-cloud applications

- Follow the Server Core pattern

Nano Server - Next Step in the Cloud Journey

Server Core

Server with Local Admin Tools

Basic Client Experience

Nano Server

Nano Server - Roles & Features

- Zero-footprint model - Server Roles and Optional Features live outside of Nano

Server- Standalone packages that install like applications

- Key Roles & Features- Hyper-V, Storage (SoFS), and Clustering- Core CLR, ASP.NET 5 & PaaS

- Full Windows Server driver support- Antimalware Built-in- System Center and Apps Insight agents to

follow

Nano Server in Windows Server 2016

- An installation option, like Server Core- Not listed in Setup because image must be

customized with drivers- Separate folder on the Windows Server media

- Available since the Windows Server Technical Preview 2 released at Ignite

Installing Drivers

- For the leanest image, install just the drivers your hardware requires- Dism /Add-Driver /driver:<path>

- Nano Server includes a package of all drivers in Server Core- Dism /Add-Package /PackagePath:.\packages\Microsoft-

NanoServer-OEM-Drivers-Package.cab

- To run Nano Server as a VM install- Dism /Add-Package /PackagePath:.\packagesMicrosoft-

NanoServer-Guest-Package.cab

Deploying Nano Server

- Generate a VHD from NanoServer.wim- Download Convert-WindowsImage.ps1 from the

Script Center: https://gallery.technet.microsoft.com/scriptcenter/Convert-WindowsImageps1-0fe23a8f

- Run Convert-WindowsImage.ps1 -Sourcepath <path to wim> -VHD <path to new VHD file> –VHDformat VHD -Edition 1.

- Dism /Apply-Image

Installing Roles and Features

- Nano Server folder has a Packages sub-folder

- Dism /Add-Package /PackagePath:.\packages\<package>

- Dism /Add-Package /PackagePath:.\packages\en-us\<package>

Installing Agents and Tools on Nano Server- No MSI support in Nano Server

- Current builds of Nano Server require xcopy or custom PowerShell script

- Nano Server Installer in the works, which will provide- Install- Uninstall- Inventory- Online and offline installation support

Installing Software on Nano Server

- No MSI support in Nano Server- Current builds of Nano Server require xcopy or

custom PowerShell script- Nano Server Installer in the works, which will

provide- Install- Uninstall- Inventory- Online and offline installation support

Demo: Nano Server Deployment

Hyper-V

Change In Market Share Over The Past 5+ Years

0

10

20

30% P

oint

s

+30.6

+6.2

CY 2008 Q1

CY 2009Q3

CY 2012Q3

CY 2014Q3

ESX

Microsoft Hyper-V Server

Q1 CY2008Windows Server 2008Released

Q3 CY2009Windows Server 2008 R2Released

Q3 CY2012Windows Server 2012Released

Q3 CY2014

CURRENT

ChangeSince Hyper-V Released

MicrosoftHyper-V

Server0.0% 11.8% 25.9% 30.6% +30.6

Pts

ESX 40.0% 46.6% 51.4% 46.2% +6.2 Pts

x86 Server Virtualization ShareFor The Past 5+ Years

Source: IDC WW Quarterly Server Virtualization Tracker, December 2014. Hyper-V and ESX + vSphere shares based on percent market share among all x86 new hypervisor deployments (nonpaid and paid). x86 hypervisor shipments include those sold on new servers, new nonpaid hypervisor deployments aboard new servers, and new hypervisor sales and nonpaid hypervisor deployments on installed base servers. Share gains for Hyper-V and ESX + vSphere come in part from market share transfers from older products from same vendors.

We are winningvirtualization share

Windows Server 2008 Released

Windows Server 2008 R2 Released

Windows Server 2012 Released

Security

Challenges in protecting high-value assets

Any seized or infected host administrators can access guest virtual machines

Impossible to identify legitimate hosts without a hardware based verification

Tenants VMs are exposed to storage and network attacks while unencrypted

Fabric

Hypervisor

Customer

Hypervisor

Fabric

Storage

Host OS

Customer

Guest VM

Legitimate host?

Guest VM

Confidently protect sensitive customer data: Designed for ‘zero-trust’ environments

Host Guardian ServiceEnabler to run Shielded Virtual Machines on a legitimate host in the fabric

Shielded VMBitlocker enabled VM

Virtual Secure ModeProcess and Memory access protection from the host

Any seized or infected host administrators can access guest virtual machines

Impossible to identify legitimate hosts without a hardware based verification

Tenants VMs are exposed to storage and network attacks while unencrypted

Hardware-rooted technologies to separate the guest operating system from host administrators

Guarded fabric to identify legitimate hosts and certify them to run shielded tenant Generation 2 VMs

Virtualized trusted platform module (vTPM) support to encrypt virtual machines

Host Guardian Service

Fabric

Hypervisor

Customer

Hypervisor

Fabric

Storage

Host OS

Customer

Guest VM

Trust the host

Guest VM

Shielded VMs

Host Guardian Service

Storage

HOST without TPM (generic host)

Virtual hard disk

HOST with TPM

Virtual hard disk

Virtual hard disk

Shielded Virtual Machines

Shielded Virtual Machines

Shielded Virtual Machines

Spotlight capabilities

Shielded Virtual Machines can only run in fabrics that are designated as owners of that virtual machine

Shielded Virtual Machines will need to be encrypted (by BitLocker or other means) in order to ensure that only the designated owners can run this virtual machine

You can convert a running Generation 2 virtual machine into a Shielded Virtual Machine

Secure Boot Support for Linux

- Providing kernel code integrity protections for Linux guest operating systems.

- Works with:• Ubuntu 14.04 and later• SUSE Linux Enterprise Server 12

Secure Boot Support for Linux

- PowerShell to enable this:Set-VMFirmware "Ubuntu" -SecureBootTemplate MicrosoftUEFICertificateAuthority

Resiliency & Availability

Storage and Cluster Resiliency

Virtual Machine Storage Resiliency

- Storage fabric outage no longer means that virtual machine crash

• Virtual machines pause and resume automatically in response to storage fabric problems

Virtual Machine Cluster Resiliency

- VMs continue to run even when a node falls out of cluster membership

- Resiliency to transient failures- Repeat offenders are “quarantined”

Shared VHDX

Improved Shared VDHX

- Host Based Backup of Shared VHDX- Online Resize of Shared VHDX

Replica Support for Hot Add of VHDX

- When you add a new virtual hard disk to a virtual machine that is being replicated – it is automatically added to the not-replicated set. This set can be updated online.

Set-VMReplication "VMName" -ReplicatedDisks (Get-VMHardDiskDrive "VMName")

Runtime Memory Resize

- Dynamic memory is great, but more can be done.

- For Windows Server 2016 guests, you can now increase and decrease the memory assigned to virtual machines while they are running.

Hot add / remove of network adapters- Network adapters can be added and removed

from Generation 2 virtual machines while they are running.

Servicing & Upgrades

Rolling Cluster Upgrade

- You can now upgrade a 2012 R2 Hyper-V cluster to Windows Server 2016 with:

• No new hardware• No downtime• The ability to roll-back safely if needed

New VM Upgrade Process

- Windows Server 2016:• Hyper-V will not automatically upgrade virtual

machines• Upgrading a virtual machine is a manual

operation that is separate from upgrading the host

• Individual virtual machines can be moved back to earlier versions, until they have been manually upgraded

New VM Upgrade Process

- Windows Server 2016:

- PowerShell only:Update-VMConfigurationVersion

Changing how we handle VM servicing- Windows Server 2016:• VM drivers (integration services) updated when

needed• Require latest available VM drivers for that guest

operating system• Drivers delivered directly to the guest operating

system via Windows Update

Scale Improvements

Evolving Hyper-V Backup

- New architecture to improve reliability, scale and performance.

• Decoupling backing up virtual machines from backing up the underlying storage.

• No longer dependent on hardware snapshots for core backup functionality, but still able to take advantage of hardware capabilities when they are present.

Built in change tracking for Backup

- Most Hyper-V backup solutions today implement kernel level file system filters in order to gain efficiency.

• Makes it hard for backup partners to update to newer versions of Windows

• Increases the complexity of Hyper-V deployments

- Efficient change tracking for backup is now part of the platform

VM Configuration Changes

- New virtual machine configuration file• Binary format for efficient performance at scale• Resilient logging for changes- New file extensions• .VMCX and .VMRS

Operations

Production Checkpoints

- Delivers the same Checkpoint experience that you had in Windows Server 2012 R2 – but now fully supported for Production Environments

• Uses VSS instead of Saved State to create checkpoint

• Restoring a checkpoint is just like restoring a system backup

PowerShell Direct to Guest OS

- You can now script PowerShell in the Guest OS directly from the Host OS- No need to configure PowerShell Remoting- Or even have network connectivity- Still need to have guest credentials

Network Adapter Identification

- You can name individual network adapters in the virtual machine settings – and see the same name inside the guest operating system.

- PowerShell in host:

- PowerShell in guest:

Add-VMNetworkAdapter -VMName “TestVM" -SwitchName "Virtual Switch" -Name "Fred" -Passthru | Set-VMNetworkAdapter -DeviceNaming on

Get-NetAdapterAdvancedProperty | ?{$_.DisplayName -eq "Hyper-V Network Adapter Name"} | select Name, DisplayValue

ReFS Accelerated VHDX Operations

- Taking advantage of an intelligent file system for:• Instant fixed disk creation• Instant disk merge operations

Isolation

Distributed Storage QoS

- Windows Server 2016:- Leveraging Scale Out File Server to allow you to:• Define IOPs reserves for important virtual hard

disks• Define a IOPs reserve and limit that is shared by

a group of virtual machines / virtual hard disks

Host Resource Protection

- Dynamically identify virtual machines that are not “playing well” and reduce their resource allocation.

Networking

Software Defined Networking

- Bringing Software Defined Networking to the next level- VXLAN and NVGRE support- Virtual Firewall- Software Load Balancer- Improved Gateways- RDMA using vNICs- And much much more…

One more thing….

Containers

- LXC (Linux Containers) is an operating-system-level virtualization environment for running multiple isolated Linux systems (containers) on a single Linux control host. Containers provide operating system-level virtualization through a virtual environment that has its own process and network space, instead of creating a full-fledged virtual machine.

What are Containers

Bare-Metal

Operating System

Physical Host

Image

Virtual Machines

Operating System

Operating System

Image

Virtual machine(s)

Operating System

Hardware Virtualization

Container Run-time

Operating SystemContainer(s)

Operating System

Hardware Virtualization

Container Ecosystem

Container Images

Container Run-Time

Linux

Image Repository

Applications

ApplicationFrameworks

Containers

Physical Machine

Hyper-V Virtual Machine(s)

Hyper-VContainer(s)

Windows ServerContainer(s)

Deploying Containers

Physical Machine

Hyper-V Virtual Machine(s)

Hyper-VContainer(s)

Windows ServerContainer(s)

Requires Hyper-V Hypervisor

Hyper-V Hypervisor

Deploying Containers

Physical Machine

Virtual Machines

Hyper-VContainer(s)

Windows ServerContainer(s)

Deploying Containers

Physical Machine

Virtual Machines

Hyper-V Container(s)

Windows Server Container(s)

Requires Hyper-V Hypervisor

Hyper-V Hypervisor

Demo: Containers

Software Defined StorageThe (v)Next Level

Current SDS with Server 2012 R2 (1 / 2)

Storage Pool

Storage Space Virtual Disks

Scale-Out File Server \\FileServer\Share

Cluster Shared Volumes C:\ClusterStorage

SMB3.x

Storage Node Storage Node Storage Node Storage Node

Soft

war

e D

efine

d St

orag

e Sy

stem

Scale-Out File ServerAccess point for Hyper-VScale-out data accessData access resiliency

Cluster Shared VolumesSingle consistent namespaceFast failover

Storage SpacesStorage poolingVirtual disksData Resiliency

Hardware- Standard volume hardware- Fast and efficient networking- Shared storage enclosures- SAS SSD- SAS HDD

Shared JBOD Storage

Current SDS with Server 2012 R2 (2 / 2)

Storage Space

HDD TierCold Data

SSD Tier and WBC

Hot Data

SAS SSD

SAS HDD

I/O Activity Accumulates

Heat at Sub-File Granularity

Hyper-V Nodes

Tiered Spaces leverage file system intelligenceFile system measures data activity at sub-file granularityHeat follows filesAdmin-controlled file pinning is possible

Data movementAutomated promotion of hot data to SSD tierConfigurable scheduled task

Write-Back Cache (WBC)Helps smooth effects of write burstsUses a small amount of SSD capacityIO to SSD bypass WBCLarge IO bypass WBC

ComplementaryTogether, WBC and the SSD tier address data’s short-term and long-term performance needs

What we hear from customers…..

I can’t implement a Microsoft Storage Solution because…

“Replication on Storage-Level is missing”

“I wanna do Hyper-Converged”

“I don’t trust Microsoft doing Storage”

Storage Spaces Direct

Storage Spaces Direct at a glance• Non-Shared Storage (inbox or local attached)• Distributed via new Software Storage Bus• Enables easy scale out• Leverages all the benefits of SMB3.x / RDMA

Storage Spaces Direct logical View

Reliability, scalability, flexibility • Fault tolerance to disk, enclosure, node failures• Scale pools to large number of drives• Simple and fine grained expansion

Use cases• Hyper-V IaaS storage• Storage for backup and replication targets• Hyper-converged (compute and storage together)• Converged (compute and storage separate)

Cloud design points and management• Standard servers with local storage• New device types such as SATA and NVMe SSD• Deploy/manage/monitor with SCVMM, SCOM & PowerShell

Scenarios

SCALE-OUT FILE SERVER CLUSTER

HYPER-V CLUSTER

Hyper-converged Converged (Disaggregated)

HYPER-V CLUSTER(S)

SMB3 STORAGE NETWORK FABRIC

- Compute and Storage resources together- Compute and Storage scale and are managed together- Typically small to medium sized scale-out deployments

- Compute and Storage resources separate- Compute and Storage scale and are managed independently- Typically larger scale-out deployments

•Cache scoped to local machine•Read and Write cache (see table below)•Automatic configuration when enabling S2D

• Special partition on each caching device• Leaves 32GB for pool and virtual disks

metadata• Round robin binding of SSD to HDD

Software Storage Bus Cache

SATASSD

SATASSD

SATAHDD

SATAHDD

SATAHDD

SATAHDD

SATAHDD

SATAHDD

Storage Configuration Caching devices Capacity devices Caching behavior

SATA SSD + SATA HDD All SATA SSD All SATA HDD Read + Write

NVMe SSD + SATA HDD All NVMe SSD All SATA HDD Read + Write

NVMe SSD + SATA SSD All NVMe SSD All SATA SSD Write only

Caching Devices

Capacity Devices

•Virtual storage bus spanning all nodes• Clusport: Initiator (virtual HBA)• ClusBlft: Target (virtual disk /

enclosures)

•SMB3/SMB Direct transport

•Intelligent I/O Management• Prioritizing(App vs System)• De-randomization of random IO• Drives sequential IO pattern on

rotational media

Software Storage Bus

ClusPort

SpacePort

Virtual Disks

File System

Cluster Shared Volumes File System (CSVFS)

Application

Node 1

Block over SMB

ClusBflt

Node 2

Physical Devices

ClusPort

SpacePort

Virtual Disks

File System

ClusBflt

Physical Devices

Storage Spaces Direct: Requirements• 4 up to 16 Storage Nodes• 10 Gbps RDMA• Min. 2 SSD / 1 NMVe per Node (Cache / Journal)• Min. 4 SSD / HDD per Node (capacity)• Supported HW Model (VMs for LAB)

Storage Spaces Direct Development Partners

Quanta D51PH

HP Apollo 2000 System

Dell PowerEdge R730xdCisco UCS C3160 Rack Server

Intel® Server Board S2600WT-Based Systems Lenovo System x3650 M5

Fujitsu Primergy RX2540 M1

DemoStorage Spaces Direct

Storage Replica

SR is used for DR preparation

Storage Replica at a glance

- Volume based block-level storage replication- synchronous or asynchronous- HW agnostic (any type of source / destination

volume)- SMB3 as transport protocol- Leverages RDMA / SMB3 Encryption

Multichanneling- I/Os pre-aggregated prior to transfer- Managed via Powershell, cluadmin, ASR

Storage Replica Layer

SMB server

File system filters

CSVFS/NTFS/ReFS

VolSnap filter

BitLocker Filter

Volume Manager

Storage Replica

Partition manager

Disk driver

Disk

Storage Replica

Partition manager

Disk driver

Disk

Source Destination

SR: Streched Cluster

- Single Cluster, spanning DCs- Asymmetric Storage- Automatic Failover- Synchronous only

Streched Cluster

Datacenter A Datacenter B

SR: Cluster to Cluster

- Multiple Cluster- Manual Failover- Synchronous or Asynchronous

Cluster A

Datacenter A Datacenter B

Cluster B

SR: Server to Server

- To separate Servers- Manual Failover- Synchronous or Asynchronous

Server A

Datacenter A Datacenter B

Server B

Synchronous workflow

Applications(local or remote)

Source ServerNode (SR)

DataLog

1

t 2

Destination ServerNode (SR)

DataLog

t1 3

2

5

4

Asynchronous workflow

Applications(local or remote)

Source ServerNode (SR)

DataLog

1

t 2

Destination ServerNode (SR)

DataLog

t1 5

4

3

6

The async difference is the IO completes on the source before travelling to the destination

Storage Replica: Requirements• Any volume (SAS, SAN, iSCSI, Local)• <5 ms round trip between sites for synchron

mirror• RDMA• Identical size for source / target volume• SSDs for log disks recommended (min. 8GB size)• Identical physical disk geometry (phys. sector

size)• Turn on SR write ordering for distributed app data

DemoStorage Replica

ReFS (Resilient File System)

ReFS - Data IntegrityMetadata ChecksumsChecksums protect all filesystem metadata

User Data ChecksumsOptional checksums protect file data

Checksum Verification• Occurs on every read of checksum-protected

data• During periodic background scrubbing

Healing of Detected Corruption• Healing occurs as soon as corruption is detected• Healthy version retrieved from Spaces’ alternate

versions (i.e. mirrors or parity data), if available• ReFS uses the healthy version to automatically

have Storage Spaces repair the corruption

Disk 1

Disk 2

Disk 3

Storage Spaces 3-way mirror with ReFS

• Checksums verified on reads• On checksum mismatch, mirrors are consulted• Good copies used to heal bad mirror

ReFS - Resiliency and AvailabilityAvailability• ReFS designed to stay online and keep your data accessible when all else fails• Even when corruptions cannot be healed via Storage Spaces’ resiliency

(mirror- and parity-generated data versions)

Online Repair• CHKDSK-Like repairs performed without taking the volume offline• No downtime due to repair operations!

On-Volume Backups of Critical Metadata• Backups of critical metadata are automatically maintained on the volume• Online repair process consults backups if checksum-based repair fails• Provides additional protection for volume-critical metadata, in addition that

provided by mirror and parity spaces

ReFS - Speed and Efficiency Efficient VM Checkpoints and BackupVHD(X) checkpoints cleaned up without physical data copies• Data migrated between parent and child VHD(X)

files as a ReFS metadata operation• Reduction of I/O to disk• Increased speed

• Reduces impact of checkpoint clean-up to foreground workloads

Accelerated Fixed VHD(X) CreationFixed VHD(X) files zeroed with just a metadata operation• Impact of zeroing I/O to foreground workloads

eliminated• Decreases VM deployment time

Quick Dynamic VHD(X) ExpansionDynamic VHD(X) files zeroed with a metadata operation• Impact of zeroing I/O for expanded regions

eliminated• Reduces latency spike for foreground workloads

Stable Point-In-Time

Virtual Hard Disk

Parent VHDX File

Child VHDX File

Hyper-V Checkpoint Merging

VHDX File Merged VHDX

Incoming Writes

VHDX on NTFS

VHDX on ReFS

Hyper-V VHDX Creation / Extension

Storage QoS

Mitigate noisy neighbor issues

Monitor end to end storage performance

Deploy at high density with confidence

Distributed Storage QoS

Storage Quality of Service (QoS)Control and monitor storage performance

Management • System Center VMM and Ops Manager • PowerShell built-in for Hyper-V and SoFS

Simple out of box behavior • Enabled by default for Scale Out File Server • Automatic metrics per VHD, VM, Host, Volume• Includes normalized IOPs and latency

Flexible and customizable policies • Policy per VHD, VM, service, or tenant• Define minimum and maximum IOPs• Fair distribution within policy

SCALE OUT FILE SERVERSERVER CLUSTER

HYPER-V CLUSTER

VIRTUAL MACHINES

Policy Manager

Rate Limiter

IO Scheduler

Storage QoS Policy Types

VIRTUAL MACHINES VIRTUAL MACHINES

Single Instance

Thank you!