Saturday, January 20, 2007

What is Disaster Recovery?

Business continuity planning

Disaster recovery is the ability to restore the business activity after the total loss of data, caused by natural disasters such as fire or flood. Data may also be lost due to other different factors including power failures, equipment failures, terrorist ttacks, deliberate disruptions and human error. Quick recovering from such disasters is a business priority and this is where a well-documented plan of recovery is neccessary. Research shows that some companies spend even up to 25% of their budget on disaster recovery plans. It is better to spend more money for disaster recovery plans than later incur bigger losses. It is said that 43% of companies that had lost all of the data never reopen, 51% close within two years and only 6% will survive (Cummings, Haag & McCubbrey 2005.)

One should pay attention to the following points when creating the plan of disaster recovery:

1. Notify all key personnel about the disaster and assign them tasks according to the recovery plan.

2. Recall backups - If backup tapes are taken offsite, these need to be recalled. If using remote backup services, a network connection to the remote backup location (or the Internet) will be required.

3. Notify all clients about the disaster - this minimizes panic.

4. Employees should work longer, more stressful hours, and a support system should be in place to alleviate some of the stress.

5. Backups (business information) should be stored in different, separate location from the company. This greatly reduces risk of data loss in case of fire for example.

Creating the plan

The strategy one should take when creating the plan of recovery depends on many factors, for example how many critical applications there are in the comapny, costs the company may spend on disaster recovery, departments that are required for proper functioning of the company, number of locations the company has etc.

These factors cause that there is no one ideal plan of disaster recovery which anticipates all possibilities and is suitable for everyone. One should think globally, as if it were a part of a larger plan of recovery. Then when a disaster strikes such a plan will be more flexible and one will be able to choose a certain option from avaiable modules that will be appropriate in a given situation. In some situations one will need help of the other company.

Another question to answer is who should take part in creation of the disaster recovery plan. Generally it is the IT department that actively commits to the strategy creation. IT proffesionals will prepare the plan with respect to its technical part. However IT department should not be the only one taking part in the prepatation of the project. The whole company should be committed to this activity. It is important to define what exactly the company risks and what the requirements are based on the following points:

1. Time to recover - the time from when a disaster strikes to the time when the business must be running properly

2. Costs of downtime - potential costs caused by the downtime and business continuity recovery

3. Point of recovery - place in time to which lost data are to be recovered, for example the beginning of a day

Data, storage, loss and recovery

Next factor one should take care of is the very data. Disaster recovery planning requires a backup application, backup media and backup records that can be taken off site, a backup schedule, and a plan for testing the recovery procedure periodically. The idea of disaster recovery is to ensure that backups are made on a routine basis and that the backups are taken offsite, so in case of a disaster that wipes out the computer center, the computers can be replaced and the data restored from the offsite backups. Lots of companies employ people just for doing backups, testing them and making sure the data are safe and secure. It is also wise to always have two backup copies in case one copy fails. The other factor to consider is to have redundant power supplies, redundant network connections etc. Redundancy greatly improves recovery process and also minimizes risk of downtime.

Testing the plan

To be well prepared when a disaster strikes you have to test, test and once again test. Only by testing will you be able to quickly recover from the disaster and get back to business. You should test at least once a year different scenarios that may happen to you. You should regularly check your backups, make sure everything is ok. Testing also helps to find out weaknesses in your plan and to correct eventual errors before it is to late.

Tuesday, January 16, 2007

Data replication for a disaster recovery

IT requirements concerning a disaster recovery are usually quite complex but if you take your time and create a disaster recovery plan then implementing these solutions will be easy and will let you be well prepared in case of an eventual disaster.

Nowadays there are a lot of companies that can help you in preparing a disaster recovery plan and train your employees how to quickly get back to business after a disaster happens. Not every company however opts for this kind of solution. Many of them just doesn’t want outside people to look at their business. It may also be quite challenging when other people tell you that your solution doesn’t work and needs changes.

Most if not all of the companies these days require a good backup solution that provides a copy of your critical business data. Companies also would like to have a guarantee that all the data the business has will be recovered in case of a disaster.

Therefore to provide companies with their needs there are several solutions available that will make the data safe and easy to recover.

The most common disaster recovery solutions that companies use nowadays are:

1. Mirroring – this solution creates an exact copy of your existing data. To provide a good recovery solution this requires a separate disk on which the original data will be stored. You would also want this separate disk to be in a different location than the original data. Having your backup disk in a different location also requires high bandwidth connection between two locations.

2. Host Based Replication – this kind of replication resides on the server that needs to have its data backed up. This solution has an advantage that the cost can be very low. However depending on how many servers the company has this cost can go up and can be significant. The downside is that it is decentralized solution and most of today’s companies trend toward centralization. The other disadvantage is that using host-based replication has impact on host’s CPU, memory and network. There are many cases that this disaster recovery solution fits in. It may be used for example for a low-cost disaster recovery protection of file servers and print servers the company has or protecting key applications the company runs.

3. Appliance-based replication – appliance is a dedicated hardware and software system designed for solving a specified task. In this case the appliance is designed for performing all replication tasks and resides between the host and the storage. The advantage is that there is no impact on host’s CPU and the host even doesn’t know that there is the appliance that takes care of backing the data. The downside is that the appliance itself may become a bottleneck if there are a lot of hosts needing replication services.

4. Storage-based replication – This kind of replication is a combination of host-based and appliance-based solutions. There is no overhead on servers and a replication takes place without them ever knowing. The other benefit is that the management is centralized and any server that the storage system supports can utilize replication features of the storage device. The company may also use solutions provided by disaster recovery vendors which use their own proprietary software and provide the company with a constant monitoring.

All these data recovery solutions have advantages and disadvantages. The solution which your company will use is dependent on several different factors. Of which the most important are: the size of your company, how critical your data is and how much money your company spends on a disaster recovery.

Sunday, January 14, 2007

Disaster recovery articles

I found a nice and promising site about disaster recovery, business continuity and disaster recovery planning. I think you will find some interesting articles there.

Here are links to these articles:

Disaster recovery planning - interesting article on disaster recovery planning and the steps that need to be taken to be well prepared in case there's a disaster.

Preparing for a disaster recovery - another article on how to prepare for a disaster recovery

Disaster recovery in San Francisco - article about a company in San Francisco that may help you in disaster recovery

Disaster recovery and your computer - list of questions on data recovery which you should answer when preparing for disaster recovery

Configuration Management Process, Roles, and Responsibilities - article on how to manage your IT infrastructure to provide business continuity