1. Microsoft ® Exchange Server 2010 Mailbox Resiliency Name Title Microsoft Corporation

2. Email Trends “Business users report that they currently spend 19 percent of their work days, or close to two hours per day, on email.” – Messaging & Collaboration – Business User Survey 2007, Radicati “The average corporate user, today, can expect to send and receive about 156 messages a day, and this number is expected to grow to about 233 messages a day by 2012. An increase of 33 percent over the four-year period.” – Messaging & Collaboration – Business User Survey 2008, Radicati Email volume is growing Users expect larger corporate mailboxes Email is business critical: Time loss after a failure is measured in seconds Data loss after a failure needs to be close to zero

3. Large Mailbox Benefits Improve user productivity − Access to all email from all clients − Less time spent managing mailbox quota − Eliminate lost or corrupted.PST files Reduce IT operations costs − Simplify email discovery and retention management − Eliminate proliferation of.PST files stored outside of IT control − Utilize high-capacity disk drives efficiently

4. Exchange 2010 Mailbox Resiliency Enables deployment of large, low-cost mailboxes Single solution for High Availability, Disaster Recovery, and Backup − Simplified administration reduces complexity − Built-in features for mailbox recovery Improved availability Storage flexibility

5. New unified solution for High Availability, Disaster Recovery, and Backup Mailbox Resiliency New unified solution for High Availability, Disaster Recovery, and Backup Evolution of Continuous Replication technology Provides full redundancy of Exchange roles on as few as two servers Reduce backup frequency through up to 16 replicas of each database Can be deployed on a range of storage options DB1 DB3 DB2 DB4 DB5 Recover quickly from disk and database failures DB1 DB2 DB4 DB5 DB3 DB1 DB2 DB4 DB5 DB3 Replicate databases to remote datacenter San JoseNew York

6. AD site: Dallas Clients connect via CAS servers DB2 DB3 DB1 DB4 DB5 DB1 DB2 DB3 DB4 DB5 DB1 DB2 DB3 DB4 DB5 DB1 DB3 DB5 DB1 AD site: San Jose Failover managed within Exchange Easy to stretch across sites Database - centric failover Mailbox Resiliency Overview

7. Mailbox Resiliency Components Database Availability Group (DAG) Mailbox Servers Mailbox Database Copies Active Manager RPC Client Access Service Active Manager Client DB2 DB1 DB2 DB3 DB1 DB2 DB3 DB1 Active Manager RPC Client Access Service DB3 AM Client Database Availability Group

8. A group of up to 16 mailbox servers that host a set of replicated databases Wraps a Windows ® Failover Cluster Defines the boundary of replication and failover Mailbox servers Host the active and passive copies of multiple mailbox databases Support up to 100 databases per server Database Availability Group (DAG)

9. Database names are unique across an forest Up to 16 copies of each database Each database has one Active copy in a DAG Database Availability Group DB2 DB1 DB2 DB3 DB1 DB2 DB3 DB1 DB3 Each server hosts only one copy of a database Replication using Log Shipping System tracks health of each copy Mailbox Resiliency Fundamentals Mailbox database copies

10. Database behind on logs (e.g Server Reboot) Database available for log replication Continuous Replication – File Mode Continuous Replication – Block Mode ESE Log Buffer Replication Log Buffer X Continuous Replication in SP1

11. Active Manager Mailbox Resiliency Fundamentals Active Manager Selects the “best” copy to activate when the active mailbox database fails 30-second database failover Process which runs on every server in DAG Provides definitive information on where a database is active and mounted − Active Directory ® is primary source for configuration information − Active Manager is primary source for changeable state information such as active and mounted Active Manager Client runs on CAS and HUB Servers

12. Double resiliency Mailbox Resiliency Design Example Double resiliency Single Site 4 Nodes in a DAG 3 Database Copies Database Availability Group (DAG) DB2 DB3 DB5 DB4 DB7 DB8 DB1 DB2 DB3 DB4 Mailbox Server 1 DB5DB6 DB7 DB8DB1DB2 Mailbox Server 2 Mailbox Server 3 DB3DB4 DB5 DB6 DB7 DB8 Mailbox Server 4 DB1 DB6 Upgrade server 1 Server 2 fails Server 1 upgrade is done 2 active copies die

13. Reduces cost and complexity Simplified Administration Reduces cost and complexity High Availability administration all within Exchange 2010 − Exchange Management Console for common tasks − Exchange Management Shell (PowerShell) Mailbox Databases managed at Organizational Level Same automated database failover process used for a range for failures—disk, server, network Simplified activation of Exchange 2010 services in a standby datacenter

14. Improved Administration in SP1 Additional Tools provided to simplify management − Active mailbox database redistribution − DAG Maintenance Mode − Single Copy Alert − Failover Metrics Reporting (Improved) − DAG property page supports static IP address specification

15. 1 1 2 2 3 3 Select a database View locations and status of replicated copies Take action (add copies, change master, etc.) Managing Availability

16. Creating a Database Availability Group Exchange Management Console Creating a Database Availability Group Only DAG Name required Witness Server and Directory automatically configured by Exchange 2010

17. Adding Mailbox Servers to a DAG Exchange Management Console Adding Mailbox Servers to a DAG

19. Create Database Copies Exchange Management Console Create Database Copies

21. Reduces cost and complexity Incremental Deployment Reduces cost and complexity Easy to add high availability to existing deployment High availability configuration is post-setup Database Availability Group Datacenter 1 Datacenter 2 DB2 DB3 DB1 DB2 DB3 DB1 DB2 DB3 DB1 Mailbox servers in a DAG can host other Exchange 2010 roles

22. Built-in site resilience solution Resiliency Across Datacenters Built-in site resilience solution Same deployment and management tools as High Availability in a single datacenter No stretched subnet networking requirements Improved process to prevent “Split Brain” Database Availability Group Simplified standby datacenter validation Faster datacenter switchover process Fewer resources required for datacenter resiliency No Client re-configuration required to access databases in standby datacenter

23. Service Pack 1 improvements Resiliency Across Datacenters Service Pack 1 improvements Improved support for 2 node datacenter resilient topologies − Two node DAGs can use Datacenter Activation Coordination (DAC) mode − DAC mode available to single site configurations Improved client experience for cross-site failovers − Administrator can manage Microsoft Outlook ® restarts vs. cross site connects

24. Backup Using Exchange 2010 Fast Recovery Data Retention Fast recovery Data redundancy Fast recovery Data redundancy Guaranteed item retention Past point-in-time database recovery Secondary mailbox for older data Reason for Backup Recovery Feature Exchange 2010 Feature Benefit Exchange 2010 Feature Benefit

25. Traditional Backup Support Traditional point-in-time backups useful for: − Point-in-time mailbox snapshots − Disaster Recovery with a single datacenter − Public folder backups VSS backup and restore supported at database level − Backup from active and passive copies − VSS Restore to Active only Exchange 2010 plug-in for Windows Server ® Backup − Volume level backup − Application (Exchange) level restore

26. Mailbox Database or Server failure….. Client disconnected for <30 seconds Mailbox Database or Server failure….. Client disconnected for <30 seconds Keeping users connected Higher Availability During Failures Keeping users connected Client DB2 DB3 Load Balanced Client Access Servers Client Access Server failure….. Client reconnects through another Client Access Server Client Access Server failure….. Client reconnects through another Client Access Server DB1 DB2 DB3 Mailbox Servers Database Availability Group DB1

27. Email Client Mailbox Server 1Mailbox Server 2 Client Access Server Keeping users connected Higher Availability During Moves Keeping users connected

28. Automatic protection against loss of queued email due to hardware failure Improved Transport Resiliency Automatic protection against loss of queued email due to hardware failure Servers keep “shadow copies” of items until they are delivered to the next hop

29. New options with performance enhancements Storage Improvements New options with performance enhancements 90% reduction in IOPS from Exchange Server 2003 Smoother IO patterns Resilience against corruption Storage Area Network (SAN) Direct Attached w/ SAS Disks JBOD (RAID-less) SATA Disks Choose from a range of storage technologies to reduce costs without sacrificing system availability Exchange 2010 storage enhancements

30. 4x increase in number of mailboxes per disk Improved Storage Utilization 4x increase in number of mailboxes per disk +500 125 Reduced IO enables more mailboxes per disk Utilize more disk capacity before performance limits are reached

31. Reduces the number of disks required RAID-less Storage Now an Option Reduces the number of disks required Just a Bunch of Disks (JBOD) configuration − One disk per database/log Database copies to provide resilience from disk failures Automatic page repair improves resiliency DB1-Active Database Log DB1-CopyA Database Page1Page1 Page2Page2 Page3Page3 DB1-CopyB Database Page1Page1 Page2Page2 Page3Page3 Page1Page1 Page2Page2 Page3Page3

32. Large Mailboxes at Low Cost Exchange 2010 no longer needs expensive, high-performing disks Use storage efficiently and lower storage costs − Larger, slower, lower-cost disks—SATA (Tier 2) Disks Maintain reliability and performance − Built-in Exchange 2010 Mailbox Resiliency features Exchange 2010 supports a range of storage options − Direct Attached Storage (DAS), Storage Area Network (SAN) − RAID and RAID-less (JBOD)

33. Mailbox Resiliency Summary

34. © 2010 Microsoft Corporation. All rights reserved. Microsoft, Exchange ActiveSync, Outlook, 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.

35. Storage Changes Deep Dive

36. Exchange 2010 Storage Changes IO reductionSATA/Tier 2 disk optimizationRAID-less storage (JBOD)Storage design flexibility

37. Exchange 2010 IOPS Trend +90% Reduction!

38. Exchange 2010 Store Improvements Improved performance Outlook Web App/Outlook online mode Outlook cached mode/Exchange ActiveSync ® Server management—Move mailbox, indexing Optimized for large mailboxes (10GB+) and 100,000 items per folder Fewer, sequential, large size disk IOs Fewer, sequential, large size disk IOs Many, random, small disk IOs Many, random, small disk IOs

39. IOPS Reductions Exchange 2010 Store Changes IOPS Reductions Sequential, large size data access Messages stored in per-mailbox database tables Database space allocated in contiguous manner On-demand view updates Increased IO size from 8KB to 32KB

40. Smoother IO Write Patterns Exchange 2010 throttles database writes Less disk contention during write bursts for a better client user experience Improved performance for SATA (Tier 2 class) disks Latency (ms) Maximum DB Write IOs Issued

41. Additional Mailbox Resiliency Slides

42. DAGS do all that SCC can, and much more Why Don't We Support SCC? DAGS do all that SCC can, and much more SCC out of the box provides little Exchange HA value − Does not cover storage failures (data is your most valuable asset) − Does not cover site failures − No failover on Store failure − Redundant network only for Clustering’s internal needs − Only provides protection from server hardware failures and bluescreens SCC has a SAN hard requirement; DAGs have no hard requirement around storage type Third-party replication API