By Jerome M Wendt DCIG, LLC
The pressures to implement business continuity software that can span the enterprise and recover application servers grow with each passing day. Disasters come in every form and shape from regional disasters (earthquakes, floods, lightning strikes) to terrorist attacks to brown-outs to someone accidently unplugging the wrong server.
Adding to the complexity, the number of application servers and virtual machines are on the rise and IT headcounts are flat or shrinking. Despite these real-world situations, companies often still buy business continuity software that is based on centralized or stand-alone computing models that everyone started abandoning over a decade ago.
Distributed computing is now almost universally used for hosting mission critical applications in all companies. However business continuity software that can easily recover and restore data in distributed environments is still based on 10 year old models. This puts businesses in a situation when they end up purchasing business continuity software that can only recover a subset of their application data.
Organizations now need a new set of criteria that accounts for the complexities of distributed systems environments. Today’s business continuity software must be truly enterprise and distributed in its design. Here are 10 features that companies now need to identify when selecting business continuity software so it meets the needs of their enterprise distributed environment:
- Heterogeneous server and storage support. In distributed environments, companies generally have multiple operating systems and storage systems from multiple different hardware vendors. Companies want the flexibility to recover applications running on any of these operating systems while using storage that they have available at the DR site to do the recovery. Many business continuity solutions require the same configurations (host software, network appliance, storage system) at the production and DR sites. New enterprise business continuity intended for distributed environments should not.
- Accounts for differences in performance. A major reason that companies implement specific business continuity solutions for specific applications is due to how they manage high numbers of write I/Os. High performance (i.e. high write I/Os) applications put much different demands on business continuity software than those that protect application servers with infrequent write I/Os. To scale in enterprise distributed environments, the business continuity software needs to provide options to scale under either type of application load.
- Manages replication over WAN links. Replicating all production data to the target site is great until the network connection becomes congested or breaks. Enterprise business continuity needs to monitor these WAN connections, provide logical starting and stopping points if the connection is interrupted and resume replication without loosing data or negatively impacting the application which it is protecting.
- Multiple ways to replicate data. Not every application server needs all of its data replicated. Some application servers need only select files or directories replicated while other application servers need all data on one or more volumes replicated to ensure the recoverability of the system. Enterprise business continuity software should give companies the flexibility to replicate data at whatever layer – block or file – that the application server requires.
- Application integration. Replicating data without any knowledge of what application is using the data or how it is using the data represents a substantial risk when it comes time to recover the application. Recovering applications such as Microsoft Exchange, SharePoint or SQL Server that keep multiple files open at the same time can result in inconsistent and unrecoverable copies of data at the DR site. Business continuity software must integrate with these applications such that it provides consistently recoverable images at the DR site.
- Provides multiple recovery points. A problem with a number of existing business continuity solutions is that it only provides one recovery point – the one right before the disaster occurred. However disasters are rarely ever that neat and tidy. Sometimes companies are not even aware a disaster has occurred until hours after the disaster (think database corruption or wrong file loaded). Business continuity software needs to provide multiple recovery points so companies can rollback to a point in time right before the disaster occurred as well as give them multiple options to recover the data.
- Introduces little or no overhead on the host server. Putting agents on host servers provides a number of intangible benefits – application awareness, capture of all write I/Os, and even a choice as to where the replication (block or file) of the data will occur. However if using agent on the server consumes so many resources on the server that the application can not run, it negates the point of using the business continuity software in the first place.
- Replicates data at different points in the network (host, network or storage system). Getting agents on every corporate server is usually never an option. Whether it is because of corporate service level agreements (SLAs), ignorance about the presence of new virtual machines or just good old fashioned corporate politics, agents are great but not an option in every situation. In this case, the business continuity software should also provide options to capture data at either the network or storage system level.
- Centrally managed. Enterprise business continuity software needs to monitor and manage where it is installed in the enterprise, what applications it is protecting, how much data it is replicating and the flow of replication from the production to DR sites. It also should provide a console from which administrators can manage recoveries anywhere in the enterprise.
- Scales to manage replication for tens, hundreds or even thousands of servers. Enterprise companies sometimes fail to realize just how many application servers they actually have in their organization. Tens of servers is a given in even most small organizations with hundreds or even thousands of servers more common than not in any large company. The business continuity software should include an architecture that scales to account for this number of servers without breaking the replication processes or the bank.
InMage is exhibiting at Storage Expo 2008 the UK’s definitive event for data storage, information and content management. Now in its 8th year, the show features a comprehensive FREE education programme and over 100 exhibitors at the National Hall, Olympia, London from 15 - 16 October 2008 www.storage-expo.com