ASR – Scenario and Workload

deep dive

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
 Failover

Hot Swap UPS Cooling

Redundancy
IT challenges implementing business continuity
Business continuity and data protection are critical issues for every organization

Cost Complexity Compliance

Datacenter cost Multiple datacenters Need to retain data

Resource cost Restoring tape Need to provide service
Hardware cost Managing management software Challenging to comply

© Microsoft Corporation
How Microsoft Azure can help
Accelerate your business continuity strategy

Reduced Reduced Increased

Cost Complexity Compliance

© Microsoft Corporation
Reduced cost
Azure benefits
Reduced
Cost

No need to purchase additional hardware

No secondary site resource costs
Reduced
Pay for what you use Complexity

Increased
Compliance

© Microsoft Corporation
Reduced complexity
Azure benefits
Reduced
Cost

Faster onboarding with cloud services

Simpler execution for testing and failover
Reduced
Integrated business continuity as a service Complexity

Increased
Compliance

© Microsoft Corporation
Increased compliance
Azure benefits
Reduced
Cost
Take advantage of Azure’s industry-
leading certification portfolio
Deploy in one of Azure’s 38 datacenters
located around the world Reduced
Increase your coverage of applications Complexity
to meet your compliance requirements

Increased
Compliance

© Microsoft Corporation
Business continuity strategy
You need all three

Primary site Primary site Secondary site Original Backup

High availability Disaster recovery Backup

When your applications When your applications have a When your data is corrupted,
have a catastrophic failure, catastrophic failure, run them in deleted or lost, you can restore it
run a second instance Azure or a secondary datacenter

© Microsoft Corporation
Resiliency in the public cloud
Resiliency is a joint effort between customers and service providers

Customer’s responsibility

IaaS PaaS SaaS Resiliency considerations

Database / data High availability, DR, backup

Workload / application High availability, DR, backup

Virtual Machine / OS High availability, DR, backup

Storage High availability, DR, backup

Networking High availability, DR

Power / facility High availability, DR

Service provider’s responsibility

© Microsoft Corporation
Resiliency in Azure
Azure provides resiliency as a platform and solutions through globe’s largest datacenter footprint

Resiliency solutions customers can use

IaaS PaaS SaaS Resiliency services in Azure

Database / data Azure Backup, Azure SQL/MySQL Database

Workload / application Azure Backup, Azure Site Recovery

Virtual Machine / OS Availability Set. Azure Site Recovery, Azure Backup

Storage Local/Zone/Geo redundant storage, Managed Disk

Networking Region Pairs, Availability Zones, IP/Load Balancers

Power / facility Region Pairs, Availability Zones, Availability Set

Azure resiliency as a platform

© Microsoft Corporation
Azure compute resiliency solutions
Apply autoscaling to virtual machines for high availability

Virtual Machine / OS VM Scale set (VMSS)

Load balancer standard

VNet
Reliably deploy and update at a large scale
Deploy hundreds of identical virtual machines in minutes.

Scale automatically
Use only the compute resources your application needs at any time. VM scale set VM scale set VM scale set
Zone 1 Zone 2 Zone 3
Simplify networking
Easily spread your workloads across the virtual machines in your
Virtual Machine Scale Set. Load balancer standard

Support hyperscale workloads VNet

Elastic to support your scale-out workloads—including stateless web
front ends, container orchestration, and microservices clusters.

VM scale set (coming soon)

Zone 1 Zone 2 Zone 3
© Microsoft Corporation
Azure application resiliency solutions
Ensure application availability with cloud-based disaster recovery

Workload / application

Source environment (East US) Target environment (Central US)

storageaccount storageaccountcacheasr storageaccountasr
Site Recovery
Data flow Replicates workloads running
Disks Disks
on Azure virtual machines
Disks Disks Cache data
(VMs) from a primary site to a
secondary location.

Data When an outage occurs at

flow your primary site, you fail over
to the secondary location and
access apps from there. After
Azure Virtual Machine Azure Virtual Machine
Failover the primary location is running
Site recovery Site recovery
ready again, you can fail back to it.
extension extension
mobility service mobility service

Availability set Availability set

Subnet1 Subnet1

VNet VNet

© Microsoft Corporation
Business
continuity
in action

© Microsoft Corporation
 Web Tier

Linux Linux
Apache Apache

App Tier

PHP PHP

Db Tier

MySQL

Hypervisor

VMware ESX

On-Premises (London)

© Microsoft Corporation
 Web Tier Web Tier

Linux Linux Linux Linux

Apache Apache Apache Apache

App Tier App Tier

PHP PHP PHP PHP

Db Tier Db Tier

MySQL MySQL

Hypervisor Hypervisor
Azure
VMware ESX Azure

On-Premises (London) Public Cloud (North Europe)

© Microsoft Corporation
 Web Tier Web Tier

Linux Linux Linux Linux

Apache Apache Apache Apache

App Tier App Tier

PHP PHP PHP PHP

Db Tier Db Tier

MySQL MySQL

Hypervisor Hypervisor
Azure
VMware ESX Azure

On-Premises (London) Public Cloud (North Europe)

© Microsoft Corporation
 Web Tier

Linux Linux
Apache Apache

App Tier

PHP PHP

Db Tier

MySQL Azure
Backup

Hypervisor
Azure
VMware ESX

On-Premises (London) Public Cloud (North Europe)

© Microsoft Corporation
 Web Tier

Linux Linux
Apache Apache

App Tier

PHP PHP

Db Tier

MySQL Azure
Backup

Hypervisor
Azure
VMware ESX

On-Premises (London) Public Cloud (North Europe)

© Microsoft Corporation
 Web Tier Web Tier

Linux Linux Linux Linux

Apache Apache Apache Apache

App Tier App Tier

PHP PHP PHP PHP

Db Tier Db Tier

MySQL MySQL Azure

Backup

Hypervisor Hypervisor
Azure
VMware ESX Azure

On-Premises (London) Public Cloud (North Europe)

© Microsoft Corporation
 Web Tier Web Tier

Linux Linux Linux Linux

Apache Apache Apache Apache

App Tier App Tier

PHP PHP PHP PHP

Db Tier Db Tier

MySQL MySQL

Hypervisor Hypervisor
Azure
Azure Azure

Public Cloud (North Europe) Public Cloud (West Europe)

© Microsoft Corporation
Why Business Continuity/DR?
Loss of Data &
Cost of Disruption Brand & Equity Compliance
Service

Organizations Loss of Clients want to

Average cost of
experience 4+ reputation is partner with
the disruption is
disruptions each often orgs that have
$1.5M/hour
year irreparable BCDR in place

4 in 10 Customer trust Protection &

Recovery times businesses do and brand recovery key
range from 1hr not reopen after severely tenet in most
to 9hrs a major disaster impacted compliances

Downtime = Loss of business to competition or closure

* Source: EMC Digital Universe with Research and Analysis by IDC, 2014
Source: IDC: Measuring Cost of Downtime and Recovery Objectives Among U.S. Firms, IDC QuickPoll Survey July 2013 and Storage User Demand Study 2013
Source: Federal Emergency Management Agency (FEMA)
Azure Site Recovery
Workload support in Azure Site Recovery

Workload / application Azure Site Recovery

Replicate Azure Replicate Hyper-V VMs Replicate Hyper-V Replicate VMware VMs Replicate VMware
Workload VMs to Azure to a secondary site VMs to Azure to a secondary site VMs to Azure
Active Directory, DNS Y Y Y Y Y
Web apps (IIS, SQL) Y Y Y Y Y
System Center Operations Manager Y Y Y Y Y
Sharepoint Y Y Y Y Y
SAP Y Y Y Y Y
Replicate SAP site to Azure for non-cluster (tested by Microsoft) (tested by Microsoft) (tested by Microsoft) (tested by Microsoft) (tested by Microsoft)
Exchange (non-DAG) Y Y Y Y Y
Remote Desktop/VDI Y Y Y Y Y
Y Y Y Y Y
Linux (operating system and apps)
(tested by Microsoft) (tested by Microsoft) (tested by Microsoft) (tested by Microsoft) (tested by Microsoft)
Dynamics AX Y Y Y Y Y
Y Y Y Y Y
Oracle
(tested by Microsoft) (tested by Microsoft) (tested by Microsoft) (tested by Microsoft) (tested by Microsoft)
Windows File Server Y Y Y Y Y
Citrix XenApp and XenDesktop Y N/A Y N/A Y

© Microsoft Corporation
Most common cause of DR is NOT natural
disasters
Any service disruption can have a major impact
Availability
 Backup is all about Data Availability
 Value Prop – Restore data
 Solution – Azure Backup
 Hero Scenarios
 Oops Delete
 Long term retention

 Disaster Recovery is all about Application Availability

 Value Prop – Recover Application
 Solution – Azure Site Recovery
 Hero Scenarios
 Disaster Recovery
 Migration

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
DRaaS Market Projection
The DRaaS market is estimated to grow from $1.42 Billion in 2015 to
$11.92 Billion in 2020, at a Compound Annual Growth Rate
(CAGR) of 52.9% from 2015 to 2020. In regional segmentation,
North America is expected to be the largest market in terms of
market size while Latin America, Asia-Pacific (APAC), and the
Middle East and Africa (MEA) are expected to emerge rapidly in
this market at high CAGRs
Source: MarketsandMarkets, June 2015
http://www.marketsandmarkets.com/Market-Reports/recovery-as-a-service-market-962.html

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
Business Continuity Challenges
Need to reduce the costs related to downtime
Costs Disaster recovery solutions with synchronous replication are expensive

Monitoring Constant monitoring of services can be challenging

Manual recovery of the many virtual machines that compose services can be
Recovery complex and time-consuming - procedures need to be documented and
tested

Protecting Many Some workloads that could benefit from protection go unprotected due to
Workloads costs and complexity

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
Safeguard your customer’s applications
On-premises to on-premises protection with Azure Site Recovery

Microsoft Azure Microsoft Azure

Site Recovery Site Recovery

Communication Channel

vCenter / vCenter
Physical

Replication channel: Replication Channel

Windows Primary Recovery Windows VMware/ Primary Recovery

Server Site Site Server Physical Site Site VMware

Support for heterogeneous environments

Key features include: Automated VM protection & replication Customizable recovery plans
No-impact recovery plan testing
Remote health monitoring Integration with existing investments
M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY Orchestrated recovery of tiered applications
Safeguard your customer’s applications
On-premises to Microsoft Azure protection with Azure Site Recovery

Microsoft Azure Microsoft Azure

Site Recovery Site Recovery

Orchestration Orchestration
& Replication & Replication

vCenter /
Physical
GA

Primary Windows Primary VMware/

Site Server Site Physical

Support for heterogeneous environments

Use Azure as your DR site Remote health monitoring
Key features include: Automated VM protection & replication Customizable recovery plans
No-impact recovery plan testing
Orchestrated recovery of tiered applications
M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
Offering DR as a Add-on Service for Hosted Workloads

DR Orchestration DR Orchestration
Microsoft Azure
Site Recovery

Extensible Data
Channel

HSP Data Center 1 HSP Data Center 2

Offering DR as a Add-on Service for Hosted Workloads

DR Orchestration Microsoft Azure DR Orchestration

Site Recovery

Extensible Data
Channel

HSP Data Center 1 Microsoft Azure

Offering DRaaS for customer workloads

DR Orchestration DR Orchestration
Microsoft Azure
Site Recovery

Customer 1 DC Extensible Data

Channel

HSP Data Center

Customer 2 DC
Offering DRaaS for customer
workloads
DR Orchestration DR Orchestration
Microsoft Azure
Site Recovery

Customer 1 DC Extensible Data

Channel

Microsoft Azure
Customer 2 DC
 Microsoft Azure Pack Integration

DR Orchestration Microsoft Azure DR Orchestration

Site Recovery

Azure Azure
Pack, Pack,
SCVMM SCVMM

Source: Hyper-V Workloads & &

Extensible Data Target: Hyper-V Workloads
DRP DRP
Channel

HSP Data Center 1 / CPS HSP Data Center 2 / CPS

Stamp 1 Stamp 2
More details Azure Site Recovery integration with WAP and Azure blog
 DR between two Microsoft Azure Pack / CPS
Stamps
DR Orchestration Microsoft Azure DR Orchestration
Site Recovery

Azure Azure
Pack, Pack,
SCVMM SCVMM

Source: Hyper-V Workloads & &

Extensible Data Target: Hyper-V Workloads
DRP DRP
Channel

HSP Data Center 1 / CPS HSP Data Center 2 / CPS

Stamp 1 Stamp 2
More details Azure Site Recovery integration with WAP and Azure blog
 DR from Microsoft Azure Pack/CPS DR to Azure

DR Orchestration Microsoft Azure DR Orchestration

Site Recovery

Azure
Pack, Microsoft Public
SCVMM

Source: Hyper-V Workloads & Azure

Extensible Data
DRP
Channel

HSP Datacenter /CPS Microsoft Public Azure

Stamp
More details Azure Site Recovery integration with WAP and Azure blog
CPS standard (La Jolla) DR to Azure

DR Orchestration Microsoft Azure DR Orchestration

Site Recovery

Azure
Pack, Microsoft Public
SCVMM

Source: Hyper-V Workloads &

Azure
Extensible Data
DRP
Channel

La Jolla Stamp Microsoft Public Azure

All configuration and Setup are automated. Every VM deployed on CPS standard is protected
Offering DR service
through Tenant
Portal/Hoster Console
ASR APIs

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
 ASR – Service Provider FAQ
 Does this solution works for dedicated or shared infrastructure model?
 ASR supports both dedicated as well as shared infrastructure models. However, CSP mandates that each azure service be
mapped to a unique tenant and therefore requires an ASR vault per customer
 .
 Will my Tenants get bill from Azure?
 No Service Providers will get ASR bill from Microsoft and they will generate Tenant specific bills.

 Will Tenant’s application data go to the public cloud?

 For Service Provider as target site - application data never goes to Azure. It is always sent encrypted over network link
between two data centers.
 For Azure as target site – application data goes to Azure encrypted over https. Here also we support public internet, or VPN
or express route.
 For Azure as DR target do we need to run the VMs in Azure all the time?
 No ASR is designed as first class public cloud DR solutions where during steady state you don’t need to run your VMs in
Azure. In case of failovers like DR drills or real failovers, ASR automatically re-creates the VMs in Azure
 Do we ensure Tenant level isolation when using Azure as target site?
 Yes ASR supports Tenant level isolation when using Azure as target site. In CSP, tenant level isolation is guaranteed by the
way CSP is constructed

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
ASR enables great
workload DR

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
Differentiate with Unmatched App Support

Windows Server (IIS, RDC/VDI, File Server)

Only Disaster Recovery Solution backed by Microsoft Support for Microsoft Applications

For SA Customers: Zero additional license charge for DR of 1st party workloads

VSS integration, App consistent, Multi-VM, in-guest orchestration

ASR Hero Scenario – Typical 3 Tier Web App

on
Azure Site Recovery DR
s trati Or
e ch
rch est
D RO rat
io n

SQL
ASR Replication
Backend
Azure Storage
App Front End
ASR Replication
App Tier

Active ASR Replication

Directory

Primary Site Azure

Above shows DR to Azure however same Hero scenario is applicable for On Premise to On Premise DR.
ASR Hero Scenario - with AD Replication and SQL Availability
Group
on
Azure Site Recovery DR
s trati Or
e ch
rch est
D RO rat
io n

SQL IaaS VM
SQL Availability Group
Backend

App Tier ASR Replication Azure

App Front End Storage

Active AD & ASR Replication IaaS VM

Directory

Primary Site S2S VPN Azure

Above shows DR to Azure however same Hero scenario is applicable for On Premise to On Premise DR.
Disaster recovery for Sharepoint using SQL Always on + ASR

tion
Azure Site Recovery DR
tra Or
es ch
O rch est
DR rat
io n

Azure
SQL virtual
SQL Availability Group
Backend machine
SQL AG
listener
SharePoint app tier ASR Replication Azure
SharePoint web tier Storage

AD / DNS AD replication AD / DNS

Site to Site VPN Azure network

Primary Site Azure

On Prem to Azure DR - FAQs
 Azure does not support VHDX. Does ASR works with
VHDX?
 Yes. ASR converts VHDX to VHD for failover. On failback, VM continues to use VHDX.

 Does ASR supports Gen2 VMs?

 Yes. ASR converts a Gen-2 VM to Gen1 for failover. On failback, VM continues to use Gen2
capabilities.
 What are the limits on disk size?
 Each disk up to 1TB (1023 GB). And one VM can have maximum of 31 VHDDs So about 32 TB VM can
be supported. Now even 64 disks VMs are available with Azure.
 Does ASR works with Linux VMs?
 ASR supports all the Operation systems supported in Azure, which includes most flavors of Windows and
Linux

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
On Prem to Azure DR - FAQs
 What about network bandwidth?
 ASR supports working with WAN Optimizer Riverbed as well as Express route.
 Can I control the network traffic used for replication?
 For Hyper-V to Azure - Yes you can use the network throttling on the host MARS agent to control how much network
bandwidth is used for replication.
 For VMware to Azure – You can leverage Windows QoS on Process Server Machine to control traffic for replication.

 Can I send replication traffic over ExpressRoute?

 Yes. You can use Express Route to send your replication traffic in a secure and fast manner. However its not mandatory.
See here
 Is it mandatory to have S2S VPN for ASR?
 Its not mandatory to have S2S VPN we work over simple internet connectivity also. However if you have S2S we also
work with that. A detailed blog on DR to Azure networking is available here

 How do you connect to VMs post failover to Azure?

 You should enable RDP for your VMs On Premises. Then post failover if you have a S2S VPN connection you can
connect to them in the same way. Or if you want to access them over internet – we have a script available with ASR
integration with Azure automation to enable RDP endpoint.
M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
On Prem to Azure DR - FAQs
 Can I retain IP after failover to Azure?
 Yes. ASR allows you to specify a IP for failover VM.
 How do I route client requests to failover Application?
 ASR works with Azure Traffic Manager to help clients requests routed to failover application. Recovery plan scripting (Using Azure
automation) can also be used for DNS updates.

 Can I failover a partial site to Azure and connect back to OnPrem site?
 Yes. Using Site to Site VPN you can connect failed over Application in Azure back to On-Premise components.

 Do I pay for running VMs for ASR protection.

 ASR is designed for Public Cloud DR. Therefore in steady state we replicate changes to storage and you don’t need to pay for any
Azure IAAS VM charges (Big competitive advantage). When you failover – ASR automatically creates IAAS VMs and then you pay
for IAAS charges for the period you run VMs in Azure.
 My OnPrem VM has multiple drives C:, D:, E: etc. When VM failover to
Azure how do I retain these drive letters?
 Please set your SAN Policy (For Hyper-V workloads) as ON inside Guest On Premise and then Azure will retain the drive letters. See
here . For VMWare workloads you don’t need to do anything we retain the drive letters for you.
 Is Exchange supported for DR to Azure?
 Yes Azure IAAS team has published Exchange support.
M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
On Prem to Azure DR - FAQs
 What Recovery Time Objective (RTO) ASR supports?
 ASR supports minutes of RTO which is industry standard. Actual failover time depends on type of VMs. Best way
to measure is run TFO and use ASR Job reports to view where time is spent as ASR provides time taken for each
activity.
 What Recovery Point Objective (RPO) ASR supports?
 ASR supports near sync RPO (In seconds) for Azure target (Best in public cloud DR). Actual RPO depends upon
the data change rate and available bandwidth to send this data to Azure.
 Does ASR supports failback?
 Yes – you can failback to on-premises.

 Is there a tool to help me plan capacity

 Yes we have an excel sheet (See RDP Kit) as well as Capacity Planner tool.

 What are the limits for ASR?

 As a customer you can protect hundreds/thousands of VMs for DR to Azure. However within a recovery plan we
support a maximum of 50 VMs. As customers use Recovery plans for Applications and have different recovery plans
for different applications.
 Each Azure subscription comes with a set of limits on cores etc. Use TFOs to validate if you need to increase the
M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
default values.
References –
 Getting Started with On-premises to Azure DR
 Getting started for Branch office to Azure DR
 Getting Started for VMware to Azure
 Getting Started with On-premises to On-premises DR using Replic
a
 Getting Started with On-premises to On-premises DR using SAN r
eplication
 Networking white-paper
 Monitoring and Troubleshooting
 Site Recovery Forum
M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY
© 2015 Microsoft Corporation. All rights reserved. Microsoft, Windows, and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the
part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

M I C R O S O F T C O N F I D E N T I A L – I N T E R N A L O N LY