Keep on Truckin’
Clive Longbottom, Service Director, Quocirca Ltd
Business Continuity – The art (or science) of maximising the chances of an organisation being able to continue its functions during and after the failure of any physical attribute of the business • This is not just an IT issue – it also has to cover buildings, people, etc.
© 2008 Quocirca Ltd
Coming up with a BC plan • Think what can go wrong – What’s PROBABLE – What’s POSSIBLE • Think of what would be needed to minimise the impact • Check viability – This changes the most • Cost out comprehensively – Not just direct cost, but indirect human costs, brand costs, etc – Just the BC costs • The DR team will do their bit
© 2008 Quocirca Ltd
Sliding Scale BC • Retaining 100% capability is not really viable • These days, a range of possibilities can be provided – Cost for 90% capability – Cost for 50% capability – Cost for 20% capability • Such approach provides flexibility and more information to the business so that an informed BC/DR agreed plan can be constructed • What are the main downsides for each %? © 2008 Quocirca Ltd
Risk Assessment/Management • BC is all about Risk Assessment and Management • The cost of managing the risk against the cost of surviving the risk has to be taken in to account – How much to ensure that the risk has minimum impact on the business – How much to just cope with whatever impact the risk brings • The cost of Disaster Recovery is something that has to be decided after the BC team has done its work © 2008 Quocirca Ltd
Common Storage Architecture Local Storage
Branch Storage
Appliance Storage Central Storage © 2008 Quocirca Ltd
Backup Storage
Issues • • • • • • • • •
Too dispersed Too many points of failure High cost of management High security risks Underutilised All that power usage All that real estate Where is that piece of data? How does it refer to that piece of data?
© 2008 Quocirca Ltd
Basic BC storage approaches • Hot data mirror – Live storage copy in-room/oncampus/off-site • Spinning virtual application images – “Hot” images ready to take over any failed application instances/platforms – Multi-instance applications have lower downtime issues – Images can be rapidly brought on line to replace broken images • Fast restore snapshots – Not quite full BC, but a lot faster than tape-based backup-restore DR © 2008 Quocirca Ltd
Scenario: Business Data • Information is held in CRM and ERP enterprise systems, as well as eMail and individual productivity tools • Customer details are replicated all over the place – Often with errors • Failure of any storage asset could have a high impact downstream to the business • Identification of problems and remediation is difficult
© 2008 Quocirca Ltd
Approach for storage 1. 2. 3. 4.
Identify your data Cleanse your data De-dupe your data Centralise your storage 5. Virtualise your storage 6. Manage your data
© 2008 Quocirca Ltd
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Where is it? • What is it? • Is it important?
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Is J. Smith the same as J. Simth? • Is 32 High St. the same as 32 Hugh St.? • Is Cat Num 12345 the same as Cat. Num. 12345?
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Use a Master Data Model • Minimise overlap – Is J. Smith the same as John Smith? – Is 32 High St. the same as 32 High Street? – Is Cat Num 12345 the same as Big Blue Cardigan?
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Compare data assets at a bit level • Get rid of files that are duplicated in different places • Minimise overlap and redundancy – and gain greater control over your information asets
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Bring it to somewhere you can manage – SAN – Maybe NAS – but only in a fully managed environment
• Use shared folders and client synchronisation for offline/mobile/remote users
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Make the many physical assets one logical pool • Make the one pool many logical units
Approach for storage 1. Identify your data 2. Cleanse your data 3. Rationalise your data 4. De-dupe your data 5. Centralise your storage 6. Virtualise your storage 7. Manage your data © 2008 Quocirca Ltd
• Use vaulting on email • Use full, in-depth reporting • Use intelligence to head off issues
And finally…. REPEAT APPORPRIATE STEPS ON A REGULAR BASIS
© 2008 Quocirca Ltd
Backup and Restore for BC • Historically, backup and restore has only been seen as a DR issue • Use of snap shots and hot backups can create a mirrorlike capability • If a backup/restore approach can bring functionality back to play before any appreciable impact on the business, then it can be regarded as BC, rather than DR
© 2008 Quocirca Ltd
Virtualisation • Traditional Wintel DAS is utilised around 10-20% – Mainframe storage generally around 60%+ • Traditional shared storage is utilised around 30% • By centralising and virtualising, utilisation can be driven to 70%+ • Virtualised storage can be used in different ways – E.g. Disk as tape – more rapid backups
© 2008 Quocirca Ltd
Tiered storage • Newest, most expensive, most performant storage is Tier 1 • Cascade previous Tier 1 storage to be on-line, Tier 2 storage • Previous Tier 2 can be used as near-line storage – Or for specific function • Cascading assets as necessary creates the optimised environment and provides for a BC platform
© 2008 Quocirca Ltd
Example: Physical Main Storage • A main storage asset could be lost. – Is the storage mirrored on a 11 basis? • No real performance hit • Cost = x2, plus management, real estate, power, etc. – Is the storage virtualised? • Incremental performance hit • Cost = +%; only for the data overhead, plus incremental management, real estate, power, etc. © 2008 Quocirca Ltd
Conclusions • Business Continuity has to be looked at in isolation of any Disaster Recovery work • Ultimate BC is out of the financial reach of the vast majority • Risk assessment and management is the key • Too much data is replicated within an organisation – Data de-duping, cleansing, MDM and centralisation are key targets • Virtualisation can really help in providing a BC platform • When BC fails, DR must be able to take over
© 2008 Quocirca Ltd