Please let others know how useful this tip is via the rating scale at the end of it. Do you have a useful Exchange or Outlook tip, timesaver or workaround to share? Submit it to our tip contest and you could win a prize.
In spite of the fact that Exchange is a mature and stable product, there is one negative aspect that you just can't ignore: it has an insatiable appetite for disk space. It is therefore important to plan for the day when your current disk resources are no longer enough. More importantly, you need to have some idea of when that day is coming.
Disk capacity planning is especially important if you are running Exchange Server 2003 Standard Edition, since it only has a 16 GB store size limit. If the server hits its 16 GB threshold, you will have to go through a rather painful recovery procedure to shrink the database and bring the server back online.
Exchange Server 2003 Enterprise Edition doesn't have a database size limit, but it's still important to do capacity planning. Sure, you might have 250 GB of free disk space, but according to Microsoft, limitations in current hardware give Exchange databases a practical limit of around 35 GB. After that, performance starts to drop.
One way of projecting future disk space consumption is to monitor the size of your Exchange database on a weekly basis. Over time, you can get a feel for how much the database is growing on average. You can then take that number and calculate the amount of time you've got until disk space becomes critical.
Unfortunately, this method is far from perfect, because your company probably does not have a static number of employees. A better way of estimating disk space consumption is to determine how much space each user consumes.
One way to accomplish this is to divide the number of bytes in the Information Store by the number of users in your organization -- but that won't give you an accurate picture because of the way that Exchange uses transaction logs.
A better way (although still not perfect) to estimate disk space consumption is to study a random sampling of users for 30 to 60 days. Microsoft recommends getting a sample size equals to at least 15% of your user base in order to get an accurate estimate. Make sure that the users you select have a variety of jobs and work in a variety of departments, so you get a sample that accurately reflects usage as a whole.
Once you have your sample group, you will want to track the number of messages those users receive on a daily basis (including spam). It's important to also note the number of messages that contain attachments, the size of each attachment and the average size of text-based messages with no attachments. Once you have tracked these statistics over a month or two, you will have enough data to make a fairly accurate projection about your server's disk consumption.
Let's suppose your data reveals that each user receives an average of 20 e-mails a day, and that those e-mails are an average of 1 KB in size. Assume as well that your study concluded that, on average, each user receives three messages with attachments and the average attachment size is 500 KB. With this information, you can do some simple math and determine how much disk space a user is consuming per day:
- 20 e-mail messages per day multiplied by 1 KB equals 20 KB
- 3 attachments per day multiplied by 500 KB equals 1,500 KB
The total amount would be 1,520 KB per user. You would then have to double your result, because Exchange stores two copies of the data -- one in the transaction log and one in the Information Store. This would mean each user is consuming an average of 3,040 KB per day.
Now, before you start multiplying this figure by the number of users and projecting how long your current disk space will last, I have to tell you that the number that you came up with is not entirely accurate -- for two reasons:
- The number assumes that the transaction logs hold onto data forever. In actuality, transaction logs tend to overwrite data after about a week, so you will have to take that into account when performing your projection.
- The formula above assumes that users keep every single message they receive. However, most users will at least delete spam.
Therefore, it might be a good idea to also look at how much mail is retained and how much mail is deleted, as a part of your study. These two factors can greatly skew your results.
About the author: Brien M. Posey, MCSE, is a Microsoft Most Valuable Professional for his work with Windows 2000 Server and IIS. Brien has served as CIO for a nationwide chain of hospitals and was once in charge of IT security for Fort Knox. As a freelance technical writer he has written for Microsoft, TechTarget, CNET, ZDNet, MSD2D, Relevant Technologies and other technology companies. You can visit Brien's personal Web site at http://www.brienposey.com.
I'm a bit surprised when you tell us that, "Transaction logs tend to overwrite data after about a week."
As soon as you make a backup, the logs are cleared as I've always learned.
Are there any good tools for measuring the message data that you mention in the last formula in your article?
I'm not sure exactly what it is you are interested in measuring, but when you install Exchange server, hundreds of Exchange-related counters are added to the Performance Monitor console. This is usually the best source for measuring anything Exchange related.
Brien M. Posey, tip author
Do you have comments on this tip? Let us know.
More information from SearchExchange.com:
- Step-by-Step Guide: How to spec your Exchange hardware needs
- Tip: Temporarily circumvent the 16 GB database size limit
- Tip: Dealing with dwindling disk space
- Tip: Beware of the Information Store's outer limits