Forefront Protection 2010 for SharePoint Performance Data

Forefront Protection 2010 for SharePoint Performance Data

Several features were added to Forefront Protection 2010 for SharePoint (FPSP) to enhance its performance capabilities. The main areas that were addressed were improving document uploads and increasing the general system performance when FPSP is added to your SharePoint environment.

In comparing the performance of FPSP against the previous version of the product, Forefront Security for SharePoint (FSSP), it was determined that there is an 81% “out of the box” improvement in uploading documents. In addition, it was estimated that adding FPSP results in an overall system performance overhead of approximately 21%.

User Model: Typically, SharePoint users navigate to FPSP for a variety of activities. During their interaction with FPSP, 15% of the time they are downloading documents, and 5% of the time they are uploading documents.  The overall upload/download rate is 2 requests per second.
Note:
This model was selected to provide a near worse-case scenario on usage that generates conservative estimates on the overall impact of adding protection to your SharePoint environment.

Content: All tests performed used a pre-chosen message mix that contained various types of Office documents, of which there was an even distribution of Office 2003 and Office 2007 data. Each file type used was comprised of varying degrees of size and complexity.

Scenarios: The following scenarios were tested:

  • Baseline: SharePoint without either the FSSP or FPSP protection solution installed.
  • FSSP OOB: FSSP SP3 out of the box product configuration (5 engines with no filters enabled).
  • FPSP OOB: FPSP out of the box product configuration (5 engines with no filters enabled).
  • FPSP Single Engine with Keyword Filtering: Microsoft Antimalware Engine only with the profanity keyword filter list enabled.
  • FPSP Single Engine without Keyword Filtering: Microsoft Antimalware Engine only with no filters enabled.

Measurements/Metrics: The measurements/metrics were as follows:

  • Upload Time (in milliseconds): The response time, in milliseconds, for end users to upload files.
  • % Impact (Uploads): The percentage impact for uploading documents over baseline.
  • % Impact (Overall): The extrapolated percentage impact of FSSP and FPSP on all SharePoint traffic based on the specified user model (including downloads and uploads).

The following table shows the test results when using the scenarios and metrics described above.

Scenarios

Upload Time
(in Milliseconds)

% Impact for Uploading Documents Over Baseline

% Impact for Overall System Performance Based on User Model Over Baseline

Baseline

538

N/A

N/A

FSSP OOB

9411

1649.26%

207%

FPSP OOB

1432

166.17%

21%

FPSP Single Engine with Keyword Filtering

847

57.43%

6%

FPSP Single Engine without Keyword Filtering

781

45.17%

3%

 

 

 

 

FPSP vs. FSSP Improvement (OOB)

 

81.83%

 

Conclusion: As previously stated, based on this testing, there is an 81% out of the box improvement in uploading document performance when using FPSP instead of FSSP. In addition, if a user performs navigation, uploads, and downloads in accordance with the 80%/5%/15% distribution (user) model, overall uploads and downloads take an estimated 21% longer with FPSP scanning than without any scanning (baseline). This number can be significantly reduced by:

  • Reducing the number of scanning engines (for example, from 5 to 1).
  • Minimizing or excluding the use of keyword filtering based on  your company's needs.  

For general tips about SharePoint 2010 performance, please refer to: http://social.technet.microsoft.com/wiki/contents/articles/7926.sharepoint-2010-tips-for-dealing-with-performance-issues.aspx

Leave a Comment
  • Please add 4 and 8 and type the answer here:
  • Post
Wiki - Revision Comment List(Revision Comment)
Comments
Page 1 of 1 (1 items)
Wikis - Comment List
Posting comments is temporarily disabled until 10:00am PST on Saturday, December 14th. Thank you for your patience.
Comments
  • Margriet Bruggeman edited Revision 10. Comment: added link

Page 1 of 1 (1 items)