The serious business of business continuity
- 30 November, 2009 22:00
In October, Air New Zealand suffered a serious failure when its mainframe went down, crashing the airline’s check-in, call centre and booking systems.
The event led to a much publicised spat with IBM, which handles the airline’s backup and disaster recovery.
In a recent survey Hitachi Data Systems discovered one-third of businesses were unable to recover files from backup tape, while a similar proportion would not be able to easily respond to a court-ordered discovery audit for emails sent and received 18 months ago.
Although Air New Zealand was able to restore its systems within six hours, the system failure highlighted that backup is not such a prominent issue. It’s recovery that matters most.
Greg Wyman, Asia Pacific regional director for StorageCraft Technologies, says remote backup was typically too complex and expensive for smaller firms, who would just use tapes.
But now businesses use systems promising Real Time Recovery, that will restore data in a matter of minutes. Best practice focuses on the Recovery Time Objective (RTO), the time it takes to rebuild a system and get it working, and the Recovery Point Objective (RPO), how long it takes to restore a “clean” data set.
Wyman says disk to disk solutions deliver best value for this, especially as businesses shift away from backup to business continuity.
StorageCraft offers products such as ShadowProtect, which allows a reseller to access a customer’s network remotely and backup data. It performs an incremental update every 15 minutes, with granular recoverability down to a single file.
Wyman says the concept of hosted solutions sounds great in theory, but in practice transferring many gigabytes of data over the net is not feasible.
Of course, WAN optimisation vendors such as Riverbed optimisation will allow much more data to be sent online.
Instead, Wyman expects offsite archiving to be the biggest growth sector in the next few years.
At Symantec, regional product manager Kerri Le says backup systems are becoming more automated, with granular recovery allowing faster access to smaller bits of required information.
Le says company platforms are now more complex through using a variety of Windows, as well as Linux, so the disks used must be able to cope with them all.
When Microsoft changes an operating system, or the price of disks comes down, this can lead to businesses re-thinking strategy, she says.
And for companies looking to switch to Windows 7, Le says Symantec can take a “snapshot” of company data just prior to the switch-over in case data is lost during the transition.
As one of the largest online server backup companies, Iron Mountain confirms a trend to full online disk-based backup, for faster restores and general recovery.
“This is further enhanced by compression and deduplication technology that requires online access to storage,” says Mac Thompson, general manager of data protection.
Customers also seek VMware support as standard, as well as the ability to offer both inhouse software solutions and a cloud SaaS solution, Thompson says.
Gen-i business manager Joe Bradley agrees, adding cloud computing promises new services.
He says Gen-i will soon offer a variety of managed services, which he brands as ‘XaaS’. These will be in addition to the company’s online storage and backup offerings.
Bradley expects enterprise customers to still seek customised solutions, but smaller outfits will go for a hybrid of customised and standard cloud-based offerings.
He says the range of technologies is increasing, virtualisation is delivering more bang for buck and disaster recovery is proving more feasible even for smaller firms.
“The biggest challenge for IT managers is making sure they have business requirements to work from and translate that into measures to build a technology solution around, such as RPO and return to operations,” he adds.
Plan-b, a company providing specialist offsite recovery facilities, says companies have faster RTOs and shorter RPOs. Virtualisation becoming mainstream means providers like Plan-b can offer cost-effective solutions, which centre around online backup and server replication.
Plan-b offers infrastructure as-a-service providing multiple ways for customers to connect to the Plan-b cloud, through online backup, replicating servers, or even from tape.
Tapes are still fine for data storage, says technical director Symon Thurlow, but difficulties in handling physical media and recovering from one server at a time is leading to unprecedented demand for online backup.
“Online backups also facilitate multiple concurrent restores of servers, reducing the RTO for multiple server recovery scenarios. We see online backup replacing daily backups in the medium term, however tape will still be used for archive,” Thurlow says.
“Replication services should not be considered a replacement for backup. If a database table is dropped or a virus infects a source server, those infections will be replicated to the replica also. Having roll back points is one way around this, but having granular backup and restore capability is key to everyday recovery also.”
Thurlow warns businesses need to consider the cost of being offline for weeks, which is why his business also offers a standby office building for staff complete with the IT infrastructure. Businesses also need to seek trusted partners, so the IT team and the management need not worry about business continuity.
“Getting a clear indication of RTO and RPO from the business is essential in order to choose what recovery technology best suits. IT are often aware of the implications of servers going down, and feel a need to do something about it. But ultimately business risk is the CFO’s or CEO’s responsibility and as such they must define the RTO and RPO requirements,” Thurlow concludes.
Revera co-founder Roger Cockayne points out there are legal requirements nowadays behind disaster recovery. Firms may also face staff simply walking away from such a disaster, so contracted providers may be necessary.
Providers can offer many solutions like data mirroring between sites, overnight data movements, disk-to-disk, or even sending a tape offsite.
Revera is phasing out tape use, but it has a role in retention for compliance needs.
“Properly using tier 1 storage infrastructure, the right enterprise software set and a team of skilled engineers just blows any other hybrid combos out of the park. Making the data safe is the first step, but only the first step for most of New Zealand’s medium and large businesses.
“The fact that data can now go into a cloud that offers full redundancy on backup and recovery, takes a leap over offerings that simply take data from one single point of failure to another.”
Cockayne says network charges are falling and customers still have control over their data with the right partner. Connectivity is improving so second sets of offices — as used by banks — are no longer needed. Customers also need to look at the contracts and guarantees in place.
Users take divergent paths to recovery
In an age of utility computing, firms like House of Travel can buy as little or as much computing power as they need.
House of Travel IT manager Dave Veronese says on-demand provisioning makes for a cost effective service, which also includes networking and firewalls.
The company recently consolidated more than 100 servers, core applications and networking, previously housed in its own Auckland and Christchurch server rooms, to Revera’s Christchurch Virtual Data Centre.
Revera likens its “rate-carded” menu-based offerings such as back-up and processing to what the sandwich maker Subway has done for lunches.
Revera’s backup systems use software to execute daily backups of specific virtual machines. Copies are housed on its storage network, addressing strict backup windows and recovery point objectives. The past 14 days of data is available online, with service levels based on data criticality determining restoration speeds.
House of Travel’s Veronese says backup is not a core competency so is best left to experts like Revera, that can plan such backups, including changing technologies or for a changing business.
“Changes are managed through Revera’s change control and SLAs. To date, the system has been flawless because Revera provides 14 days near storage, so we have never had to go back to tape other than to test,” he says.
PGG Wrightson used to use Symantec Backup Exec but claims it was slow and prevented daily backups.
The company turned to Revera and Commvault to manage its Windows-based environment with VMWare on top, aand the vendors designed a unique solution for the business.
There were challenges in handling new technologies like deduplication, but daily backups happen easily to one central location.
“Restores are very quick, since as part of the solution, we purchased a dedicated Hitachi AMS SAN, so we’re able to hold a considerable amount of online data for restores,” says PGG Wrightson IT manager Richard Kay.
PGG Wrightson is very happy with the set up and is looking to extend Commvault further into areas like storage management, compliance and archiving.
Kay adds the company may have underestimated the differences between Commvault’s offerings and traditional backups, so companies must invest in training.
For Auckland-based engineering design company URS New Zealand, the continuity challenge arrived recently in the form of a fire in its building at College Hill. The company moved into alternative premises offered by Plan-b.
URS also has a tape backup strategy using grandfather-son tape backup and offsite storage.
Off-site recovery presents a challenge though, as any system updates must be replicated at the disaster recovery (DR) centre, says Asia-Pacific IT director Shane Primrose.
Directly after the fire URS was able to operate through a functioning DR site with a WAN connection within four hours.
Primrose says URS is happy with the off-site facility, and is now looking to remove tape backups with wire backup technology, allowing quicker backups and easier management.
Disaster contingency must be planned for he says and the DR facility must be in a convenient location. As well, a DR service provider must be able to implement 90 percent of the systems with minimal input.
Meanwhile, the management at bedding company Sealy recently switched to StorageCraft’s ShadowProtect backup solution.
The company had suffered extended downtime due to server failure, with its former solution unable to restore recent backup tapes. The company’s annual IT review also discovered the solution was only writing minimal data to tape in a readable format.
Sealy chose the Disaster Recovery Group (DRG) to handle backup with DRG’s offsite remote disaster recovery solution, which is said to lose no more than 15 minutes of data in the event of a crash.
Every 15 minutes Sealy’s systems are backed up and replicated to an off-site DR datacentre. If there is a failure or outage, the company can contact DRG who can bring Sealy’s production servers up in a virtual environment, and redirect users to that virtual server in its datacentre.
The IT manager at Sealy, Jeremy Bain, says the system hasn’t let him down and he no longer has to worry about the automated system.
“I was physically handling the tapes myself. Now it just happens,” he says.