Planning Required Server Hardware

                 

 
Special Edition Using Microsoft SharePoint Portal Server
By Robert  Ferguson

Table of Contents
Chapter  6.   Capacity Planning Within Your Environment


Planning the required hardware is an essential step in the pre-deployment process. The most important hardware components to plan for are the speed of the CPU, amount of RAM, and hard disk space. SharePoint Portal Server is a CPU- intensive application. Failure to provide sufficient CPU resources will result in unacceptable server response times, which will lead to a less than satisfied user experience. CPU planning should accommodate user loads during peak and non-peak periods without impacting the users' server response time. Another important hardware component is RAM. Failure to provide sufficient RAM will also result in unacceptable server response times. Lastly, you need sufficient hard disk space to accommodate multiple document versions. Failure to plan for sufficient disk space will result in an inability to search for or save documents to the portal.

Many organizations struggle to plan for performance and scalability requirements. The challenge is that it is difficult to forecast how the site will be adopted by your end user community. Once the initial site is deployed, it is important to monitor the resources of the site and how well the current hardware specifications are handling the load. The resources to monitor are addressed later in this chapter.

Table 6.1 discusses the important deployment metrics that are needed to effectively plan your hardware for your portal environment.

Table 6.1. Planning Deployment Metrics

Deployment Metric

Description

Number of Users

Total number of potential users for a site.

Percent of active users per day

A percentage of the total users that may actually use the dashboard site during a given day. The range is 10 “100%, but is often overestimated. The average is usually approximately 30%.

Number of operations per active user per day

The number of operations that a typical user performs on the dashboard site during a given day. The number usually ranges from 1 to 10.

Number of hours per day

The total number of hours in which portal activity occurs. The range is from 12 to 24 hours.

Peak factor

An approximate number that estimates the extent at which the peak dashboard site throughput exceeds the average throughput. The number range is from 1 to 4.

NOTE

One may be able to estimate the required number of active operations per user per day by analyzing the Web server log of another portal server within the environment.


NOTE

When analyzing the Web server log, one should only review page views, not site hits. The reason is that site hits are often much higher than Web page views.


The preceding deployment metrics are used in the following formula to provide the peak throughput in operations per second:

(number of users xpercent of active users per day x number of operations per active user per day x Peak factor) / (360,000 x number of hours per day)

NOTE

Within the formula, the number 360,000 is determined by 100 (for percent conversion, so you don't have to convert the percentage to a decimal) x 60 (number of minutes in an hour ) x 60 (number of seconds in a minute).


NOTE

HTTP is the communication protocol between a client and the SharePoint Portal Server. HTTP is a connectionless protocol, and the number of concurrent users cannot be accurately measured. This is why Operations per second is an essential measurement component.


Calculating a Small Site

Let's assume that a small business unit consists of 1,200 total users. Of the 1,200 users, one particular business unit consists of approximately 300 users who use the site on a daily basis. As a result, this particular business unit represents 25% of the active users of the site. We estimate that the active users average 10 operations per day with a peak factor of 4. The hours this site is active is estimated at 12 hours per day. In addition to the 300 active users, 800 “900 users hit the site on an occasional basis. Table 6.2 below shows the deployment metrics that will be used in the calculation formula:

Table 6.2. Small Site Calculation Example

Deployment Metric

Quantity

Number of users

1,200

Percent of active users per day

25

Number of operations per active user per day

10

Number of hours per day

12

Peak factor

4

Assuming these metrics, the anticipated peak throughput is .28 operations per second.

(1200 x 25 x 10 x 4)/(360,000 x 12) = 0.28

The above site could be accommodated with a server such as a 500MHz Pentium III with 512 megabytes (MB) of RAM.

Calculating a Medium- Sized Site

Another example is a medium-sized portal consisting of a total of 5,000 users, 1,000 of which are active users. The percent of active users per day is thus 20%. We estimate that the active users will average 6 operations per day with a peak factor of 3. The hours this site is active is estimated at 14 hours per day. Table 6.3 below shows the deployment metrics that will be used in the calculation formula:

Table 6.3. Medium Size Site Calculation Example

Deployment Metric

Quantity

Number of users

5000

Percent of active users per day

20

Number of operations per active user per day

6

Number of hours per day

14

Peak factor

3

Assuming these metrics, the anticipated peak throughput is .36 operations per second.

(5000 x 20 x 6 x 3)/(360,000 x 14) = .3571431.1852

The above site could be accommodated with a server such as a 500MHz Pentium III with 1 gigabyte (GB) of RAM.

Calculating a Large Search Site

The last example is a large search site portal, consisting of approximately 20,000 users. The portal site is used to search over content that was crawled from other sources across the organization. The site is not used to manage a large number of documents, and user loads are distributed throughout the day due to time zone differences. The 20,000 users are infrequent portal site users, considering this is not the main portal site they access. We will assume the number of active users is 15% that will average 4 operations per day with a peak factor of 2. However, we will assume this site gets activity approximately 20 hours out of a 24-hour day. Table 6.4 below shows the deployment metrics that will be used in the calculation formula:

Table 6.4. Large Site Calculation Example

Deployment Metric

Quantity

Number of users

20,000

Percent of active users per day

15

Number of operations per active user per day

4

Number of hours per day

20

Peak factor

2

Assuming these metrics, the anticipated peak throughput is .33 operations per second.

(20000 x 15 x 4 x 2)/(360,000 x 20) = .333333

The above site could be accommodated with a server such as a 700MHz Pentium III with 2 gigabytes (GB) of RAM.

NOTE

It should also be noted that in order to enhance the performance of the server dedicated to searching content, it is recommended to consider a separate dedicated server for crawling content. Usually, a good time to consider a dedicated crawling server is when the search site reaches a threshold of 100,000 documents.



                 
Top


Special Edition Using Microsoft SharePoint Portal Server
Special Edition Using Microsoft SharePoint Portal Server
ISBN: 0789725703
EAN: 2147483647
Year: 2002
Pages: 286

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net