Providing Application High Availability Stephanie Balaouras Principal Analyst

Providing Application High Availability Stephanie Balaouras Principal Analyst Forrester Research

Ground Rules & Expectations • House Keeping! – All teleconference lines are open. . . – Mute yourself when not speaking and don’t put call on ‘hold’. • Collaborate! – This is a discussion, not a presentation… – Ask questions, make comments and share experiences. – And beware, Forrester may call on you! • Are we in a ‘rabbit hole’? – If the conversation has gone down a substantial ‘rabbit hole’… – … Forrester will interject, but you decide whether or not to persist.

Agenda • Understanding disaster recovery and availability • Measuring availability • Availability strategies •

Definition: disaster recovery (DR) Focuses on recovery procedures of critical IT systems following a significant disaster event or disruption that renders the primary site unavailable 4 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Definition: high availability (HA) Focuses on the technology and processes to prevent application/service outages at the primary site or in a specific IT system domain 5 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

HA and DR requirements 6 Entire contents © 2008 Forrester Research, Inc. All rights

Definition: continuous availability Combines DR preparedness and high availability strategies to ensure IT service availability whethere is a system or site failure 7 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Sources of service outages Apps Unplanned downtime Planned downtime Data integrity (data corruption, bugs) Hardware upgrades (servers, storage, networks etc. ) Data center/facilities failures (extreme weather, power outages, manmade events. ) Software upgrades (OS, DB, app) Hardware failures (servers, storage, networks etc. ) Maintenance (backups, patches) Process failures (human error, coding errors) 8 Entire contents © 2008 Forrester Research, Inc. All rights reserved. Data migrations

Availability focuses on service, not elements • Availability is the aggregation of all elements supporting the IT service. – Most components spec to 99. 999% availability – Combined IT service availability would fall below 99. 999% • Scheduled downtime is a critical part of the availability design Most downtime is planned! 10 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Availability is more than just “ 9”s • “ 9”s are misleading • Consider 99. 9% availability (8 hours/year outage for 24 x 7), What is the difference between: – 8 AM to 4 PM on the last Friday of the quarter – Biweekly outages of 30 min at 4 AM local time • Timing and duration are more important than total downtime • Well written SLAs must take all these factors into account 11 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Availability metrics • ITIL availability management – Availability of IT services compared to agreed upon SLAs – Duration of disruptions to IT services – Number of infrastructure components with availability monitoring – Number implemented measures to improve availability 12 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Local HA options Seconds FT hardware FT VMs Clustering Local standby Restart VMs Recovery time Minutes objective (RTO) System backup and restore Hours Less More Cost 14 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

IT service continuity (DR) options 15 Entire contents © 2008 Forrester Research, Inc. All

Recommendations • Downtime isn’t just the result of disasters • Make availability a part of every enterprise architecture, app development and IT infrastructure decision – The most seamless form of continuous availability is baked into the application – Preventing failures by architecture/design is cheaper than providing recovery for poorly designed applications 17 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Recommendations (cont. ) • Design applications for availability – Use reliable elements — server, storage, OS, etc. – Select HA and DR options for critical systems – Scrutinize operational process to reduce human error – Make availability a key component of acceptance testing 18 Entire contents © 2008 Forrester Research, Inc. All rights reserved.

Thank You! 19 Stephanie Balaouras Forrester Research, Inc. +1 617. 613. 6440 sbalaouras@forrester. com Megan O'Donoghue Forrester Research, Inc. +1 617/613 -6059 modonoghue@forrester. com Karen Popeo Forrester Research, Inc. +1 617/613 -6381 kpopeo@forrester. com Scott Sheehy Forrester Research, Inc. +1 617/613 -6523 ssheehy@forrester. com Entire contents © 2008 Forrester Research, Inc. All rights reserved.