|
|
DATA MANAGEMENT
Prepare for Downtime
Downtime can cause huge losses for any organization. Find out how to identify your most critical servers, and some guidelines for determining the potential costs of downtime.
It's obvious that downtime is bad for business. Therefore, you have to a strategy and solution in place in your organization to mitigate any consequences of downtime. You first have to understand which servers/business processes are critical and deserve the most protection before you can start to determine how to protect them.
A survey of the business activities of an organization usually quickly reveal the business IT applications that are critical to the smooth operation of the business. Understand that the importance and critical nature of a business process is always driven by the nature of the application, not the data.
The most common and obvious applications include:
- E-mail
- Document Management Systems
- Customer Relationship Management
- Sales Force Automation
- Support and Help Desk
- Critical business applications
- Intranet/extranet/Web applications
It's also important to recognize that the level of criticality increases with the period of downtime. For example, losing a non-critical application for a day is a nuisance, but not commercially dangerous. In some cases, even losing the data from today and resorting to last night's backup tape is acceptable.
However, with critical applications, the level of risk and probable damage increases exponentially the longer they're unavailable. One hour of downtime may be redeemable, whereas 48 hours is probably not.
The recovery time from the failure of a critical server is dependent on what went wrong, when it happened, and the steps required to deliver a fully operational solution re-connected to the users. It's almost impossible to determine the likely period of downtime for any given server, so rather than try to determine every possible eventuality and what would be required to recover, review the probable commercial risk and damage associated with a length of downtime. Keep in mind that a 60 minute outage is often as likely as a 12 hour outage.
To identify your critical servers:
- Produce a list of servers and the applications they run.
- Consider the implications of downtime of each of these AFTER a discussion with a sample of users and their manager. Good decisions are made on good information, so it is worth getting feedback.
- More than 75 percent of e-mail users concluded that loss of e-mail was more stressful than divorce in a recent survey. This feedback provides some insight into users' dependency on applications, yet most e-mail servers remain unprotected.
Here are some suggestions for calculating the likely costs of downtime:
- Detail the number of users for each application and the number of hours that they would use it in a normal day.
- Add a rough valuation of the cost per hour for that group.
- Downtime is a major variable in terms of risk and likely cost, so calculate a number of them: 15, 30, 60 minutes; 4, 8, 16, 24, and 48 hours.
- Calculate the total lost hours of productivity and the direct costs for each downtime period.
- The real cost is in the lost productivity and time sensitivity of the user and that depends on the use of the application. For example, a sales operation may go offline for a day. That is lost time that cannot be "caught up" and would represent a loss of 0.4 percent of revenue per year from this source alone.
- Put an estimate of loss against the lost productivity of those unproductive hours based on the purpose and use of the software application and user feedback.
This is very simplistic, but you'll notice that the real cost of downtime is probably much higher than you anticipated and critical applications are clearly identified.
Rather than look at and address every single server at once, focus on the most critical and address the highest area of risk.
This article is an excerpt from Neil Robertson's article in the August/September 2005 issue of SharePoint Advisor Magazine. In the full article, Neil goes on to discuss data protection, application software protection, reliability and monitoring, switch/failover criteria and performance, bandwidth considerations, and more. Subscribers can read the full article at http://Advisor.com/doc/16847.
ARTICLE INFO
Web Edition: 2006 Week 33, Doc #17912
 FREE ACCESS
Keyword Tags: Administration, collaboration, database management, Data Recovery, E-Mail, messaging, microsoft, microsoft access, microsoft office, microsoft server system, microsoft sharepoint, microsoft windows, Microsoft, Microsoft SharePoint, Microsoft SharePoint Portal Server, Microsoft Windows SharePoint Services, portals
ADVISORAMA Anyone can hold the helm when the sea is calm. -- Publilius Syrus |
|