Our solutions

First of all selecting your cloud provider. We need a large scale cloud provider who will provision and bill by the hour. Here is a table of some of the large cloud providers:

Provider

Per Hour Billing

Size

Windows Server 2008

Windows SBS 2011

Amazon

Focus per hour billing

Large

Yes

No 1

Rackspace

Mixed inc. per hour billing

Large

Yes

No

We also looked at Dell which although has a cloud offering is clearly not their main focus, HPCloud is currently in Private Beta.

Licensing

Microsoft licensing can be complex. On the Amazon cloud you can pay for Windows Server and SQL Server instances by the hour. The rental of the software is included in the per hour cost. So SQL server costs about 60c an hour or $5256 a year.

Any other licensing needs to be done through your own software licensing. If you have software assurance then your likely to be able to have DR rights for a cold backup of your systems, otherwise talk to us.

Connect

In order to get the benefit of an extensible data centre you need to be able to connect your local network to Amazon's cloud. To do this you need a router configuration which can be a Cisco or Juniper medium scale router or another routing package such as the free PFsense that supports the BGP protocol through the VPN tunnel.

Backup

In order to be able to turn the infrastructure on quickly this means that you need to backup to the cloud. Double-Take have a cloud offering which allows a pay as you go model or you can use the open source Bacula backup.

The backups are done at a granular level for each service. The list of services typically involved will be:

  • Identity (active directory)
  • File services
  • SQL Server
  • Exchange
  • Market data feeds e.g. Bloomberg, ThomsonReuters Datastream

Activate

In order to activate the data you need to be able to configure Amazon to spin up your infrastructure. We can help here with a custom application that essentially reduces the backup process to two options with a control as simple as a light switch. Turn on the disaster recovery in a test mode which gives you an isolated environment to test DR or a full scale DR activation. The main difference is what you do at the end of the activation. In a test you will discard any changes and your main systems will not be affected. In a real scenario you will actually have a migration back to either your original hardware or new hardware. We can help with this transition.

FSA Business Continuity

Disaster Recovery is just one part of the business continuity project. The FSA has published a Business Continuity Management Practice Guide which they expect firms to form a useful basis on which to structure discussions on business continuity planning. Some of the key features are A1.2.5 which is that plan should include total destructive loss of the site. From an IT point of view this is worst type of disaster and if you can cope with this then you can cope with any of the smaller disasters that will happen more frequently. It is standard practice to be able to recover with 4 hours and to be able to test every six months.

Not possible currently due to licensing restrictions by Microsoft. This may be changing as Microsoft are working out what to do with OnLive who seem to be in breach.

© 2012 Drummonds.net Ltd trading as cloudhaven.co.uk| 05759862 | VAT 888955337