Recovery Solutions for Applications Running in the Cloud

Stephen Weber BC/DR Architect & Program Manager McKesson IT Service Continuity Office (SCO)

DRJ Fall World · September 8, 2014

•  Cloud Overview –  Architecture of the Cloud

–  Architecture of Applications in a Cloud

•  Recovering Applications in the Cloud –  Define your requirements

–  Understand your “CLOUDS” (Capabilities, Limitations, Options, boUndaries, Design and Services)

–  Design, Test, Document, Implement and Certify

•  Cloud Recovery Options •  A Demonstration of Cloud Recovery

Cloud Recovery Solutions Agenda

9/8/2014 McKesson Corporation Confidential and Proprietary 2

Cloud Overview

9/8/2014 McKesson Corporation Confidential and Proprietary 3

•  Physical Layer

•  Network / SAN

•  Compute (CPUs, Memory and Drivers)

•  Storage

•  Security Appliances

•  Management and Monitoring –  Performance - Load Balancers

•  Access and Access Control

9/8/2014 McKesson Corporation Confidential and Proprietary 4

Cloud Overview Architecture of the Cloud

•  Database Servers

•  Application Servers

•  Reporting & Utility

•  Access Layer: Web, Authentication, Access Control

•  Dependencies –  Internal Environments

–  External Environments

9/8/2014 McKesson Corporation Confidential and Proprietary 5

Cloud Overview Architecture of Applications in the Cloud

Recovering Applications in the Cloud

9/8/2014 McKesson Corporation Confidential and Proprietary 6

Document & determine: •  Applications that need to be recovered •  Systems that support the applications that need to be recovered •  Detailed profiles for each of those systems (CPU, memory, storage, network, etc.)

•  Relationships, dependencies required (inputs, outputs, authentication, encryption, etc.) •  Objectives of cloud-based recovery services (RTO, RPO, levels of service, etc.) •  Access Requirements (end users, support & monitoring, connectivity, locations, etc.)

•  Performance and service level requirements •  Regulation requirements •  Backup, restore & retention requirements

9/8/2014 McKesson Corporation Confidential and Proprietary 7

Recovering Apps in the Cloud Define Recovery Requirements

Know Your

Boundaries!

9/8/2014 McKesson Corporation Confidential and Proprietary 8

Recovering Apps in the Cloud Understand Your “CLOUDS”

Capabilities Limitations Options boUndaries Design Services

•  Capabilities: –  What platforms and OSs can be supported in the cloud (AIX, Linux,

Windows, P2V , V2V, etc.)?

–  Recovery Time Achievable / Recovery Point Achievable –  Mimic production, monitoring, reporting, control, performance,

management, service –  Provide virtual clustering, firewalls, load balancers, L3 routing?

–  Are databases supported (Oracle, SQL, Exchange, SharePoint)?

–  Encryption in-flight and at rest –  Support for virtual appliances

9/8/2014 McKesson Corporation Confidential and Proprietary 9

Recovering Apps in the Cloud Understand Your “CLOUDS”

•  Limitations: –  OSs & versions, memory, CPU, storage & other (UEFI)

–  System or application recovery

–  What constitutes system recovery: multiple IPs, validation of services, drive letter mapping, volume and data validation, etc.

–  Infrastructure recovery (Active Directory, Domain Name Services, appliances, and application recovery)

–  Recovery priority order and failover groups

–  Circuit connectivity, bandwidth, VPN licenses or IP addresses

9/8/2014 McKesson Corporation Confidential and Proprietary 10

Recovering Apps in the Cloud Understand Your “CLOUDS”

•  Options: –  Continuous data points –  Full or partial failovers –  Support hybrid environments –  Flexible test time –  Scripting

•  boUndaries: –  Is there a limit to the number of servers that can be recovered to achieve

your RTO? –  Testing (same IP, sandbox, hybrid, access, failover, VPN, replication, etc.) –  Roles and responsibilities

9/8/2014 McKesson Corporation Confidential and Proprietary 11

Recovering Apps in the Cloud Understand Your “CLOUDS”

•  Design –  Requirements on the production side

–  How the data is replicated to the recovery side

–  Design requirements on the recovery side

•  Services –  IaaS, PaaS, AaaS, RaaS, etc.

–  Adjusting the level of services up or down

–  Monitoring, management and reporting

9/8/2014 McKesson Corporation Confidential and Proprietary 12

Recovering Apps in the Cloud Understand Your “CLOUDS”

•  Obtain or work with Recovery Service Provider to complete a detailed design –  Can the RSP provide a tool to gather profiles of the systems that need to be protected?

•  Test the cloud recovery service before implementing the solution in production •  Document the process as you go through the POC test •  Implement the solution:

–  Obtain a detailed implementation project plan and tailor it to your environment

–  Determine resources needed and amount of effort required for the project

–  Monitor performance levels on the replication connection

•  Certify the results: –  Conduct a failover test for each server protected

–  Conduct a final failover test with all systems

9/8/2014 McKesson Corporation Confidential and Proprietary 13

Recovering Apps in the Cloud Design, Test, Document, Implement and Certify

Cloud Recovery Options

9/8/2014 McKesson Corporation Confidential and Proprietary 14

•  Cloud Vaulting Services

•  Physical to Virtual Recovery Services (P2V)

•  Virtual to Virtual Recovery Services (V2V)

•  Hybrid Recovery Services

9/8/2014 McKesson Corporation Confidential and Proprietary 15

Cloud Recovery Options Let’s Review Four Solutions…

•  Data is backed up and replicated to a vault located at a remote location

•  Systems must be restored with OSs before data can be restored

•  RTO is typically 24 hours

•  RPO is typically 24 – 48 hours

•  Supports most platforms

•  Protection at the file level

9/8/2014 McKesson Corporation Confidential and Proprietary 16

Cloud Recovery Options Cloud Vaulting Services

•  Replication agent typically required on production server for host-based replication.

•  Systems, applications and data are replicated real-time

•  RTO within 4 hours depending on service provider and platform

•  RPO within minutes or hours, depending on service provider

•  System state and data are replicated real-time

•  Services are limited to certain platforms, OSs, and applications

•  Protection typically at the volume level, not the file level

9/8/2014 McKesson Corporation Confidential and Proprietary 17

Cloud Recovery Options Physical to Virtual Recovery Services

•  Agent or agent-less solutions available

•  Some cloud-based replication products contain their own data mover; others dependent on third party data movers

•  RPO within seconds or minutes depending on the product

•  RTO typically within 4 hours or less, depending on the Recovery Service Provider

•  Protection typically at the volume level, not file level

9/8/2014 McKesson Corporation Confidential and Proprietary 18

Cloud Recovery Options Virtual to Virtual

•  Recovery solution comprised of both cloud-based recovery solutions and traditional hot-site recovery

•  Requires infrastructure between the cloud environment and the hot-site environment

•  Network connectivity maybe limited to Layer 2 (bridging) or Layer 3 (routing)

•  Additional components might be required to provide VLAN routing

9/8/2014 McKesson Corporation Confidential and Proprietary 19

Cloud Recovery Options Hybrid Environment

9/8/2014 McKesson Corporation Confidential and Proprietary 20

Cloud Recovery Options Typical Design

Cloud Recovery Demonstration

9/8/2014 McKesson Corporation Confidential and Proprietary 21