The basics of backup

You spent big bucks on a backup solution — tape, disk-based or hybrid. You believe you're covered if disaster strikes — at least, your data is safe. But is it really? And does it even matter if your employees, partners and customers don't have a way to easily access that data after a hardware meltdown or a natural disaster? A good backup solution is an important part of an effective business continuity plan, but it's only one component.

As your company grows, it becomes more and more important to have a comprehensive strategy in place that will allow you to continue operations as normally as possible whether you lose a single hard disk, a server, or an entire building housing your network infrastructure. That's why your business continuity plan needs to be multi-layered and able to expand in scope as your company grows. There are several important issues to address when mapping out your business continuity plan — these include data backup; hardware redundancy; software failover; continuity of network connectivity and the human factor

 

Data backup
You can think of data backup as the core component of your plan. After all, hardware can be replaced, operating systems and applications can be reinstalled and new network connections can be established — eventually. But data collected or created by your users may be unique and difficult or impossible to recreate exactly.

A good data backup solution depends on the amount of data generated and how often it changes. Daily backups may suffice for small companies, whereas larger companies that produce a high volume of data may need to back up data several times a day or on a continuous basis. Tape, disk or a combination of the two, can be used for backup.

Another option is a backup service, such as Acme Data, LLC, that allows you to back up your data to a remote location across the Internet. Off-site backup is a crucial part of any backup strategy that has business continuity in mind, since backup tapes or disks stored on-site can be destroyed along with the original data if a flood, fire, tornado or other natural disaster occurs. However, data stored in a remote location may be difficult to access and restore if you lose your Internet connection. Thus, we recommend a combination of on-site and off-site backup.

An alternative for off-site backup storage is to physically move a copy of each day's backup to another location. In a small company this may mean the owner or IT manager takes the backup disk home with him/her every night. In a larger company or one with high security concerns, the backup tapes may be transported via armoured car to a vault.

 

Hardware redundancy
If a server goes down, you can lose money and productivity if you have to wait for a new machine or component to be purchased and configured. As the company (and IT budget) grows, you should plan to purchase critical servers in pairs and configure them identically at initial setup. You can then either keep the "spare" in reserve in case of failure or implement server clustering with failover so that when one of the pair goes down, the other automatically takes over with virtually no downtime.

 

Software continuity
Server clustering allows you to create a "mirrored" server that has the same software as the primary server and either mirrors the data or shares a connection to the data storage array.

If you run Windows 2000 Advanced, Datacenter Server, Windows Server 2003 Enterprise or Datacenter edition, you can use Windows' built-in clustering services. However, in Windows Clustering the cluster members (nodes) must be located near one another because they use SCSI connections to storage resources. For best protection, members of a server cluster can be located in different physical locations. Solutions from vendors such as Veritas and NSI allow mirroring of cluster servers to remote sites.

Your business continuity plan should also take into account availability of client software for accessing resources. The larger your company, the more workstations will be involved. One way to ensure that clients will have the proper software to do their jobs, even if they have to work from different machines in a different location, is to use an Application Service Provider (ASP) or have users run applications on a terminal server rather than installing the application software on individual client machines.

 

Network connectivity availability
Access to data and software may not be all that employees need to get their work done. More and more, an Internet connection is necessary to perform necessary tasks — if you run ecommerce sites your sales will depend on the availability of an Internet connection.

You can ensure continued Internet connectivity by purchasing two or more Internet connections. Software solutions such as Rainfinity's RainConnect and hardware solutions such as Xincom's Twin WAN series can be used to provide failover when one connection fails. As a bonus, many of these products also aggregate the bandwidth of your multiple Internet connections, giving you a bigger "pipe" when both connections are up and running. As the company grows, you can add more, faster connections. For example, a small company may aggregate two DSL lines or a DSL line and a cable connection. Larger companies can often use the same consolidation/failover product to aggregate multiple T-1 or T-3 lines.

Firewalls, routers and other internal network components can also be purchased with failover capabilities.

 

The human factor
Don't forget the human factor in putting together your plan. In case of a true disaster, some of your personnel may be temporarily or permanently out of commission. If the network administrator is the only one who knows the passwords that are necessary to get a critical server back up and running, time and business can be lost. Responsible password management can make emergency personnel transitions smoother.

Important passwords can be stored in a safe off-site, to which a trusted party (also off-site) has access. Cross training of IT personnel ensures that you don't run into a situation, for example, in which Joe is the only person who knows how to boot the Exchange server.

As the company grows, these practices should become more formalised and administrative responsibilities should be shared and delegated to avoid focusing all control in the hands of one person.