Plan B Disaster RecoveryDisaster Recovery Service

ZDNet Logo

This review was carried out by ZDNet in the first week of December 2008. It is reproduced below with the permission of ZDNet, and can be found on the ZDNet site at:

http://reviews.zdnet.co.uk/hardware/networking/0,1000000696,39574100,00.htm

8.0  Editors’ Rating  Excellent

Service & support 8.0
Features 7.0
Setup & ease of use 9.0
Performance 8.0

Plan B Disaster Recovery

Roger Howorth ZDNet.co.uk

Published: 08 Dec 2008

Running an IT infrastructure without a proper disaster recovery (DR) plan is reckless, given the vulnerability of IT systems and our dependence upon them. The problem is that DR plans require a mixture of services such as offsite storage, replication and networking that are tricky and expensive to put together. This is why many businesses — especially small and medium-sized ones — will be interested in a new appliance and service offering from Plan B Disaster Recovery.

Plan B, a UK company based in Berkshire, uses an on-site appliance to make daily snapshot backups of your servers, which are sent to the company’s datacentres via secure internet links. These snapshots are used to create virtualised replicas of your servers at the Plan B datacentre, and these virtual servers (Plan B calls them ‘rescue images’) are tested each day to make sure they boot properly without errors. Plan B also creates IPSec VPN links between your office and the Plan B datacentre, so the replacement servers can quickly and easily be reconnected to the company LAN. It can also run extra DNS servers to reroute traffic to the replicas in the event of the main systems going offline.

Plan B says that it can replace a client’s servers within 30 minutes of the main systems going offline. In a full-on disaster scenario where all of a client’s normal IT systems become unavailable, users can connect to the Plan B datacentre from any PC or notebook using a web browser and SSL-based VPNs. Once a suitable replacement office is available, it would be connected to the rescue images using the IPsec VPN. A three-year contract for 50GB of data will cost you £1,135 (ex. VAT) in setup fees plus a service charge of £200 a month.

Plan B’s appliance plugs into your network and takes daily snapshots of your servers, which are then sent to the company’s datacentre where virtualised ‘rescue images’ are kept.

We tested the Plan B DR service in ZDNet UK’s Labs by configuring the appliance to protect one of our servers running Windows Server 2003 Small Business Edition. Currently the appliance is compatible with all forms of Windows Server from 2003 onwards; support for a few Linux distributions is expected to be added soon.

One of big advantages of Plan B’s offering is its simplicity of setup and management. In our tests, the first step was to fill in a simple email questionnaire that briefly described the servers we wanted to protect. This helps Plan B get a handle on things like the Windows Active Directory structure, network topology and the services running on the servers. It also provides an opportunity to exclude certain directories from the snapshot backups. For example, we excluded several local disk drives that store backup data because we didn’t want to slow down the transfer of snapshots unnecessarily by moving this data around.

We also were able to specify how our internet bandwidth would be used to send data to Plan B. We took the default options, which were to limit bandwidth to 10KB/s during working hours, and to allow all our bandwidth to be used after 11pm. It took us about ten minutes to complete the questionnaire.

Next we installed the Plan B appliance, which merely involved fitting the 1U server in our datacentre rack and connecting it to mains electricity and the LAN. The initial configuration requires a keyboard and monitor to be connected to the appliance, but once the setup is complete these can be removed.

The Plan B appliance’s text-based welcome screen.

Once the appliance was switched on, we were greeted by a text-based welcome screen asking us to log into the initial configuration menu. The username and password for this were shown on the screen. Once logged in, some IP settings are required, so we told the appliance to retrieve suitable information from our DHCP server. Then we entered our customer ID and appliance ID, both of which had already been sent to us by email. Finally we typed in what Plan B calls the ‘bootstrap password’, which was imparted during a phone call with the company’s engineers. Changes can only be made to the appliance configuration by entering a bootstrap password, and a new one is needed each time a change is made, so system administrators need to contact Plan B and request a new bootstrap password before they can update the appliance configuration.

Downloading a Win32 agent onto the target server from the Plan B appliance.

The next stage is to install an agent onto the server you want to protect so that the appliance can make snapshot backups. The agents are downloaded and updated via the appliance, so the next step in our installation process was to connect to the appliance’s web-based management console from the target server. Again, Plan B had already supplied us with suitable login credentials. Once logged in, we selected the agent installation menu option and downloaded the Win32 agent. Installation onto our server was very quick, and when it was completed we used the Windows Services management tool to confirm that the Plan B service had been properly started. No reboot was required.

The agent makes backups of the servers using Microsoft Volume Shadow Copy Service, so it’s compatible with a wide range of applications and can make snapshot backups with only a momentary freeze of the Windows environment.

After enabling the agent from the Plan B appliance and ensuring that the target server’s virtual counterpart is ready at Plan B’s datacentre, snapshots can begin.

Next we used the Enable Agents option in the appliance management portal to scan the network for our newly installed agent. The scan worked like a charm, and once located we could easily enable the agent by clicking on a link. This completed all of the necessary setup at our end. However, before the appliance can begin taking snapshots, the server must be enabled at the Plan B datacentre. The Plan B operators did this as soon as they noticed our setup was active, so the first snapshot was made and transferred to the appliance almost immediately we completed our installation.

The initial snapshot is likely to contain a lot of data — in our case the server’s C drive had some 40GB. Understandably this took a few hours to transfer to the appliance, and then quite a lot longer to transfer to the Plan B datacentre. With our internet connection and bandwidth settings the first snapshot was transferred after two overnight sessions. We received a phone call from Plan B to tell us the first transfer had finished, and the engineer confirmed that our rescue image had booted properly.

We tested a disaster recovery scenario by phoning Plan B to ask them to activate our rescue image. They called back about ten minutes later with details of how to connect to the datacentre via a web browser and an SSL VPN connection. We could then connect to our virtual server and its web-based management tools using Microsoft Terminal Services. Although the rescue image had a different IP address to the original server, its applications worked perfectly. We normally mange our Labs server using LogMeIn remote control software, and this was also available and working on our virtual replica without needing to use the VPN.

Files and directories stored on the Plan B appliance can be restored to your server via the management portal’s Retrieve Data option.

Our Plan B appliance had 638GB of storage capacity, and an added bonus is that system administrators can restore individual files and directories from the appliance by logging into the web management portal. We tested this by clicking on the portal’s Retrieve Data menu option and selecting a snapshot from a list of those stored on the device. This produced a hierarchy of the files, and from here we could drill down into the directory structure and select the files to be restored. Files could be restored to their original location or to any other server running the Plan B agent.

Conclusion
We were very impressed by Plan B’s service, which we found straightforward to set up and configure. Part of the Plan B offering is to ensure that the rescue images are transferred properly every day, and the company will notify customers if this doesn’t happen properly. So once up and running we were confident we could leave it to get on with things.

As far as maintenance is concerned, the appliance receives software updates and patches automatically via its connection to the Plan B datacentre, and it distributes updates to the agent software as required — all without rebooting the servers that are being protected.

Besides the monthly service charge and setup fees, there is a disaster invocation fee of £400 to activate each virtual machine, and other charges if you don’t get your servers back online within two weeks. Plan B recently had an ISO 27001 Stage 2 audit, and expects to have this certification within a few weeks of this review’s publication.

Plan B are delighted to announce that BSI auditors have passed the organisation as meeting the requirements of the international Information Security Management Standard ISO27001:2005.

Tim Dunger, Plan B’s Operations Director said, “We are delighted that our security management structures and processes have been ISO27001 certified. We take our customers’ and our own information security very seriously and it is great to have this endorsement of our processes, people and technology.’

Plan B are the only provider of a virtualised disaster recovery service to be ISO27001 certified.

Computer Weekly yesterday reported that:

Networks were still unavailable today, nearly two days after a virus caused a “major incident” at Barts and The London NHS Trust.

Computer Weekly has learned that the virus caused a plethora of spurious messages to overload the trust’s network.

The trust, which has installed the “Cerner” Care Records Service under the National Programme for IT [NPfIT], diverted ambulances to neighbouring hospitals for several hours while IT specialists shut down the networks and tried to find out how the virus penetrated them.

The virus was discovered on Monday afternoon. By this morning there was still only limited access to the trust’s networks.

A spokesman for Barts said today that parts of the network are being brought up gradually. He said it was still not known how the virus got into the networks. It affected networks at the trust’s three main hospitals: St Bartholomew’s in the City, the Royal London in Whitechapel and the London Chest in Bethnal Green.

Barts has England’s biggest centre for treating heart attacks, and also has specialist treatments for cancer. It is one of the capital’s leading trauma and emergency care centres, and is home to London’s air ambulance.

Ambulance services to the trust were diverted during the outbreak but early evening yesterday the trust began taking in ambulances again.

Hospital officials were telling the media, on the basis of their guidance notes, that the Care Records Service - part of the £12.7bn National Programme for IT - was working normally. They said the Care Records Service was unaffected by the virus.

But one official at Barts conceded that, with the network overloaded because of the virus, staff “might” have been unable to access the Care Records Service. He said that in that event staff would have reverted to accessing paper-based records.

Although this is a normal backup procedure it can cause backlogs of work and reduce the number of patients that staff and clinicians can see and treat.

The Care Records Service allows access to patient records. The hospital’s networks also handled requests for X-rays.

In a statement yesterday the trust said, “The Trust’s well rehearsed emergency procedures have been activated to ensure that key clinical systems continue while network access is being established.

“We have maintained a safe environment for our patients throughout the incident. Manual backup systems are in use and we are in the process of restoring the computer systems with priority being given to the most important areas for maintaining patients services.

“Operating theatres and outpatients departments have remained operational throughout the incident, though some non-essential activities have been scaled back. A&E remains open to walk-in patients and ambulances are being diverted to neighbouring hospitals in the short term.”

Julian Nettel, chief executive at the trust, said yesterday, “This has been a difficult day, but by using back-up systems, manual procedures and working flexibly, we have continued to provide high quality care to our patients.”

The trust says that medical staff have been able to make paper-based requests for laboratory tests and X-rays. Nettel said, “I would particularly like to thank all our staff, patients and other NHS colleagues for their hard work, help and support during this incident.”

Plan B DR has announced a file backup enhancement to its mananged disaster recovery service.

With this lastest version of the Plan B DR service, customers can now use the snapshot data held locally on their Plan B appliance as a fast and simple file backup. If customers want to recover a lost or deleted file, they can now simply locate it on their appliance using the appliance portal to retrieve it.

We’ve also got one very large advantage over most existing backup
solutions. Because the rescue images are tested daily, the back up files are too. So you can be completely confident that your backup files will be there and work when you need them too!

More.

Plan B DR have passed the pre assessment for ISO/IEC 27001:2005 security standard with no non-conformities.