Overview:
Business Continuum, our business continuity model addresses all 4 "IT Severity Levels" of Disaster:
| Severity Level | Examples |
|---|
| 4 Displacement | Loss of Office infrastructure |
| 3 System loss | Loss of server (theft, flood, fire, outright hardware failure) |
| 2 System integrity | Loss of disk, system corruption |
| 1 Accident/Malicious | Loss of individual files through user error, deletion, incorrect data loaded |
At the heart of Business Continuum from IntraLAN Group plc is a uniquely designed and custom-built unit called the Local Data Manager (LDM). The LDM is a significantly powerful unit that is plugged anywhere on your local network where it takes a complete system image of every server, every night. It then also fully replicates itself to off-site datavaults, including entire SQL and Exchange databases, over your existing DSL line.
It achieves this by taking a complete (16bit checksum verified) copy of each server at high speed over your LAN every night. Last night's is compared with tonight's to create a perfect Differential File. This file is then compressed, encrypted and transmitted to our primary datavault where it is applied to its own copy. Following a second checksum, this datavault is now a second perfect replication of your servers. The final 2 steps are the creation of a daily "subtraction" Differential File so that up to 31 days of roll-back images are always available and then the datavault is further replicated to a second at another site over fibre.
On inception and then every 6 months, we will restore each of your servers. Unlike other DR solutions (which either don't offer complete server restores at all or provide you with x days access at their physical site with "their" hardware for you to restore yourself), we do this for you. At a pre-arranged time, we send you a link and using a pre-agreed password, you simply log on and test all your applications. On confirmation, we then re-issue you a certificate confirming your back-up status.
Ultimately, if you lost your offices (see also "FAQs"), we would perform the same restore but host your servers at our DataCentre in Canary Wharf (100Mb access). Your users can then work from home, Serviced offices anywhere with Internet access - using Terminal Services whilst you sort out whatever has happened to your offices.
The LDM:
The LDM is the cornerstone of our Business Continuum solution. Through this topology design you have:
- Instant individual file recovery (no popping home for the last tape, cataloging it and restore) it's a simple drag and drop via the LDM's web interface, all at LAN speeds,
- Even if you lose an entire server (e.g. disk failure/corruption, theft etc), with the LDM in place a complete 100Gb server still only takes c.6 hours to rebuild including the System State. Try doing that with a tape (even assuming it is fully functional)...
- Because it's local and holds a local image of each server, it can generate a Differential file between tonight and last night. Typically, a 100Gb server will lead to a 500Mb Differential file easily sent over an aDSL line overnight.
In order to appreciate this technique, consider the alternative: all files that have changed during a day's work would need to be sent in their entirety often impossible for larger files such as Exchange or SQL databases. Some solutions then try to get around this by using Incremental updates (each change is added to the previous). The problem is any fault in any one increment change makes every subsequent change invalid (say day 3 of your accounts incremental file is flawed clearly all changes since would then be invalid). Our solution always compares to the base-line and then checksumed.
- Makes long-term archives simple (especially useful for businesses operating under Compliance). Simply plug one of our pre-prepared external/portable disks into the back of the LDM to make a further "deep storage" copy as often as you like (remember you always have the last 31 days on the datavault).
Resilience:
- If there's a problem with your DSL line, the LDM still maintains a perfect local up-to-date copy. Once the line is restored, it'll simply sends a bigger Differential file to the datavault. You would need to lose your Internet line and your server(s) and the LDM all at the same time to then rely on a day-or two old copy off the datavault that's pretty robust!
- If there's a problem with the LDM itself (hardware failure/theft/damage), we'd only need to replace it (with completely blank disks). On the first night it'll simply take a full copy of all servers at LAN speeds (safe again!) and the following night restart its Differential files to the datavaults. Don't forget there's still a 24 hour old copy on the datavaults.
- If there's a problem with your server, the LDM is good to the point of server failure. Simply fix the server and restore off the LDM at local speeds.
- If there's been a corruption of your server (but it took a few days to notice), fix any hardware required, we restore off the LDM (still with corruption) but then also run the "subtraction" Differential files off the datavault to the pre-corruption state.