Business Continuity Strategy

From Supporting Role Wiki
Revision as of 13:36, 7 March 2014 by Mvanbeek (talk | contribs)
Jump to navigationJump to search

This page documents all of the procedures and processes we have in place that can be used to ensure that your business can survive a disaster, at least from an IT perspective.

Data Backups

Forget About IT Ltd currently has three backup servers in two separate locations. Every night, at approximately 10pm, our servers start to copy all the changes that have happened to your data since the previous backup. Should this backup be interrupted, it will automatically try again after a short period of time, for a number of attempts. The process is repeated for at least one other out of the three backup servers, so that your data is backup up onto at least two backup servers.

Once the backup has finished, the backup servers take a copy and date it. The daily backups are kept for a week. One backup a week is used as the weekly backup, and is kept for a month, and one backup a month is kept for a year. So at the end of a full year's rotation, we will have, on at least two different servers, at least 20 snapshots of your data (6 x daily, 4 or 5 x weekly, 12 x monthly).

Mission Critical data

Because we charge for the amount of data we store for you at our data centres, it is the client's choice as to what they consider to be mission critical data. However, unless a client specified otherwise, we back up all the data on the server.

Archive data

As all business accumulate data, eventually there comes a time when it needs to be archived off. We can provide areas on the server that are not backed up to our off-site storage, but in order to make sure the data is safe, we recommend that the client purchases a couple of external hard drives, which the backup program will use to carry out a local backup of that archive area. The client then swaps those drives on a weekly basis, taking the other off site somewhere, like their home.

Large Files

Because it is possible for files to be too large to be uploaded overnight, files over 1GB in size are only backed up at the weekend, when we can leave the backups running long enough to complete the task before the next backup is due.

Files over 10GB are not currently backed up remotely for the same reasons. However, our backup program logs what files it has not backed up, so it is very easy for us to determine if there are files on your server not being backed up due to their size.

As upload bandwidth improves over time, we review these sizes on a regular basis.

Bandwidth

we normally limit the amount of upload bandwidth we use on a client's broadband to 25% of its capacity. At the weekend we increase this to 50%. This means that when the backup program is running, it does not impact the the normal traffic to and from the Internet.

Virtual Machines

We have installed virtual computers on our servers for some of our clients. Because of the size of the image files, we do not back these up remotely. However, we set up special scripts in our backup program that can access the mission critical data, and add it to the backup.

Data Recovery

Thankfully, full metal recoveries are rare. Much more usual are accidental deletions, or using "save" instead of "save as". we can either email you the file, so that you can re-save it yourself, or we can recover the file(s) to their original location, or a new location. Obviously the larger the file, the longer it will take to restore.

Test Restores

We are happy to carry out test restores for you at any time. Should you wish to test our backup system, then I suggest you create a file and save it somewhere on the server that also contains mission critical data, and a couple of days later, delete it and ask us to recover it.

Hardware

All the hardware we use, including the servers we install at client's premises, follow a common format. These means that even when we buy servers from a different manufacturer, they remain compatible with each other. This means that in the even of a hardware failure, we can simply turn up with a spare server, swap the hard drives over, and the client is back up and running again.

If the hard drives have also been damaged, then depending of the clients preference, we can either download the backup data onto the spare server before we bring it over, or bring the server over with a base install, and then prioritise what data is downloaded first depending on the client's requirements.

We can also take the replacement server to alternative premises, along with a small network, or set it up at our office or our server farm, to be used remotely.

Mirrored Hard Drives

All of our servers used mirrored hard drives. That means that the system can tolerate a hard drive failure without any data being lost, and more importantly, no interruption to the server. If this happens to your server, we would simply turn up with a new hard drive.

Spares

As previously mentioned, we hold a number of spare servers in stock. we normally keep a 10% ratio of spare servers to production servers.

Duty of Care

The only thing we require of our clients is a duty of care over the physical well being of the server. We always use quiet servers, as they are usually located under a desk or in a corner of an office, as most of our clients do not have the space (nor do they need) for a dedicated server room. However, as such it can be exposed to spilt drinks, localised heat sources like a fan heater, etc.

Should the failure of the server be linked to one of these sorts of issues, we would normally expect to charge the client for the repair. We also expect the client to cover the cost of replacement hardware and our time in the event a major incident, such as a fire, in their insurance.

Email

Relay Servers

"Panic" Server

Security

Traffic Encryption

Data Encryption

Cloud

Remote Access