The bonding between bandwidth, backup and continuity

By Mark Lewis, EMEA marketing director, Riverbed Technology.

Organisations today are facing a growing problem when backing up data from remote sites. As companies expand, more offices are distributed across different continents or regions, linked together by public or private networks. Regardless of the size of the office or its location, the data produced by each of these sites needs to be backed up to ensure that an organisation can recover its files in the event of a disaster.

Backing up remote office data onto local tape libraries has always been an administrative and budgetary challenge. To ensure security, compliance and sidestep the challenges of maintaining a network of tape libraries, many IT teams are moving to centralised backup solutions. But the move to centralisation is not by itself a solution to the problem of backing up distributed data. While there are inherent challenges, new wide-area data services (WDS) technology can help significantly reduce the hurdles of network computing.

Companies are faced with three large challenges when it comes to the network and backups – the time taken to backup data; the time taken to restore data; and the network costs of achieving these two things.

Time to backup
Most organisations schedule backups to run during non-business hours. This creates restrictive backup windows, limited bandwidth, and network latency that can significantly reduce the probability that backup operations can complete successfully. This results in managers requesting larger backup windows that can push into the company’s working day. To compensate, organisations rotate backups of different portions of the data. If disaster were to strike, the remote site may lose many days worth of information, and at best would be down for several days whilst restoring the data – something that is just not acceptable in sectors such as banking where information is the lifeblood of the organisation.

Time to restore
In the event of a disaster or loss of data, the ability to recover files is critical. Organisations utilise the wide-area network (WAN) to transfer recovered data from the central office to the remote office. Unlike backups, which happen at a scheduled time, recovery is a live task that can halt the productivity of the workforce and the revenue of the business until it is completed.

Network costs
Data protection processes are bandwidth intensive. Backup operations, though often run at night, can overrun into business hours and compete with users for limited bandwidth. Increasing bandwidth can be expensive and often will not even improve backup or restore times. Inherent limitations in network protocols can create idle time during data transmissions no matter what the size of the pipe – slowing down the process.

To centralise backup and recovery operations and build a more secure system for protecting data while reducing costs, IT managers must ensure that they have the mechanisms in place to optimise network traffic and speed-up the process.

Complete WDS solutions accelerate the performance of all applications running on WANs over TCP by addressing the three bottlenecks to WAN performance: constrained WAN bandwidth; TCP throughput drop-off with latency; and application chattiness/latency. Efficient optimisation solutions work in harmony with each other to provide the highest levels of performance improvement and can accelerate applications up to 100 times faster in some cases.

Depending upon the product chosen, the optimisations will likely include a mechanism for reducing the amount of data that traverses the WAN, TCP window sizing modifications, application-specific optimisations to reduce the impact of application protocols such as CIFS, MAPI, SQL, HTTP, SSL and others, and high-speed TCP implementations for enabling large files to transfer across long distances. Adopting these functions is essential to deliver backups and restorations within tight time frames.

By thinking about WAN issues, it can be easier to draft a comprehensive plan to centralise data protection processes. A well-thought out design that incorporates the appropriate WDS technology will scale with a growing company and easily accommodate new sites that are installed onto the company network. A centralised model can provide the company with increasing returns on its initial hardware investment as more offices are attached to the network. And while WDS can make centralised backup and restoration processes operate as if they were on the same LAN as the remote servers, it also opens the path to an even greater goal: using WDS to deploy the server-less branch office.

http://www.riverbed.com/

•Date: 19th July 2007• Region: UK/World •Type: Article •Topic: IT continuity
Rate this article or make a comment - click here