1 2 Previous Next 11 Replies Latest reply on Feb 1, 2011 3:18 PM by brentil

    VirusScan Enterprise performance tests

    brentil

      Over the last year or so I've performed several benchmark tests to compare the performance of VSE.  I've started doing P3 tests and wanted to create a central place for all my benchmarks to make them easier to find for future use.  Feel free to add suggestions or provide results from your own benchmarks.  I plan to continue updating this thread as new updates and versions arrive.

       

      You can find benchmarks for the following items here.

       

      • CMD Line 5300 vs CMD Line 5400
      • CMD Line V1 DATs vs V2 DATs
      • 8.5i vs 8.7i
      • 8.7i 5300 vs 5400
      • 8.7i Whitelisting (added DAT circa ~5770)
      • 8.7i Gold vs P1 vs P2 vs P3 - 2010.03.19
      • 8.7i P3 vs P4 - 2011.01.06
      • CMD Line 5400 vs 6.0.3 - 2011.01.06

       

       

      Message was edited by: brentil on 1/6/11 8:59:01 AM GMT-05:00

       

       

      Message was edited by: brentil on 1/6/11 9:08:43 AM GMT-05:00
        • 1. Re: VirusScan Enterprise performance tests
          brentil

          I run each test 3 times and average the results to get the values shown. Each run was an On-Demand Scan with the Default settings + Decode  MIME, Scan Inside ZIPs, and all Unwanted Programs checked. I used  slower hardware but this is still a level of machine at the edge of what is still being used (they just have 2gb of RAM  instead). I also used the slower hardware to help show any difference  since gains would be more pronounced.  As each new test is done I will re-run the previous test as a new baseline and then run the tests of the new configuation.

           

          Machine Configuration was  as follows;

           

          Dell Dimension 4600
          P4 Hyperthreading 3.2ghz
          512mb  of RAM
          W7 RTM fully patched
          VSE 8.7 + AntiSpyware module
          Office 2007 fully patched
          VS2008 SP2  fully patched
          SQL2005 Management Suite fully patched
          FireFox,  Dreamweaver, Flash

           

          All CMD Line tests are run using these options;

          /NOMEM /MIME /unzip /PROGRAM /SUB /SECURE /clean

          • 2. Re: VirusScan Enterprise performance tests
            brentil

            Benchmarks from June 2009 done initially during the 5400 engine beta.

            • 8.5i P8 with 5300 engine w/o Artemis = 158 minutes

             

            • 8.7i with 5300 engine = 150 minutes
            • 8.7i with 5400 Beta1 engine = 148 minutes

             

            • 8.7i P1 with 5300 engine = 102
            • 8.7i P1 with 5300 engine w/o Artemis = 105
            • 8.7i P1 with 5400 Beta1 engine = 105 minutes
            • 8.7i P1 with 5400 Beta2 engine = 104 minutes
            • 8.7i P1 with 5400 engine = 101 minutes
            • 8.7i P1 with 5400 Beta2 engine (w/o Artemis, Cookies, Registry) = 88 minutes
            • 8.7i P1 with 5400 engine (w/o Artemis, Cookies, Registry) = 89 minutes

             

            • CMD Line 5300 (w/o Artemis, Cookies, Registry) = 82 minutes
            • CMD Line 5400 Beta2 V1 DATS 32bit (w/o Artemis, Cookies, Registry) = 78 minutes
            • CMD Line 5400 Beta2 V2 DATS 32bit (w/o Artemis, Cookies, Registry) = 77 minutes
            • CMD Line 5400 V2 DATS 32bit (w/o Artemis, Cookies, Registry) = 79 minutes
            • 3. Re: VirusScan Enterprise performance tests
              brentil

              These benchmarks are from October 2009 from when McAfee added Windows file Whitelisting to the DATs.  I also started using a more indepth testing process.

               

              Each test was run 3 times and averaged together. For the boot times I  used a stopwatch and observed the system until it was fully loaded into a  logged in windows state and the drive stopped thrashing. Each of those  tests was run 6 times while also not looking at the stop watch to  attempt to get a real average since it is a more subjective analysis.  The 35,000 files used was a direct copy of the Program Files folder with  the apps previously mentioned.

               

              Copy 35,000 (~6gb) files  with verification


              • No Anti-Virus = 14.3 min
              • VSE 8.7 P2 -  5400 - PreWL = 43.3 min
              • VSE 8.7 P2 - 5400 - WL = 39.3 min
              • Gain:  9.2%

               

              RAR  35,000 files


              • No  Anti-Virus = 76.3 min
              • VSE 8.7 P2 - 5400 - PreWL = 87.3 min
              • VSE  8.7 P2 - 5400 - WL = 86.7 min
              • Gain: 0.7%

               

              unRAR  35,000 files


              • No  Anti-Virus = 14.3 min
              • VSE 8.7 P2 - 5400 - PreWL = 30.7 min
              • VSE  8.7 P2 - 5400 - WL = 29.7 min
              • Gain: 3.3%

               

              System  Boot Times


              • No  Anti-Virus = 49.5 s
              • VSE 8.7 P2 - 5400 - PreWL = 56.3 s
              • VSE  8.7 P2 - 5400 - WL = 52.9 s
              • Gain: 6%
              • 4. Re: VirusScan Enterprise performance tests [VSE 8.7i P2 vs P3 testing]
                brentil

                March 2010 - VSE 8.7i P2 vs P3 testing

                 

                Engine                5400

                DAT                    5918

                McAfee Agent      4.5.0.1429

                Buffer Overflow and Access Protection DAT version: 480

                 

                Each test was run 3 times and averaged together. For the boot times  I  used a stopwatch and observed the system until it was fully loaded  into a  logged in windows state and the drive stopped thrashing. Each of  those  tests was run 6 times while also not looking at the stop watch  to  attempt to get a real average since it is a more subjective  analysis.  The 35,000 files used was a direct copy of the Program Files  folder with  the apps previously mentioned.

                 

                No big leaps forward but at least there were no real steps backwards.

                 

                Copy 35,000 (~6gb)  files  with verification


                • VSE 8.7  P2 = 42.3 min
                • VSE 8.7 P3 = 41.7 min
                • Improvement: + 1.4%

                 

                RAR  35,000 files


                • VSE 8.7  P2 = 91 min
                • VSE 8.7 P3 = 91 min
                • Improvement: 0.0%

                 

                unRAR   35,000 files


                • VSE 8.7  P2 = 30 min
                • VSE 8.7 P3 = 32 min
                • Improvement: - 7%

                Full System Scan


                • VSE 8.7  P2 = 73 min
                • VSE 8.7 P3 = 73 min
                • Improvement: 0%

                 

                System   Boot Times


                • VSE 8.7  P2 = 50 s
                • VSE 8.7 P3 = 48 s
                • Improvement: + 4%
                • 5. Re: VirusScan Enterprise performance tests

                  Hello Brentil,

                   

                  You deserve an applause and here it is...

                  • 6. Re: VirusScan Enterprise performance tests [VSE 8.7i P2 vs P3 testing]

                    Brentil,

                     

                     

                     

                    HATS OFF TO YOU ...

                     

                     

                    A great job, Very helpful and informative !!!!

                     

                    Sameer

                    • 7. Re: VirusScan Enterprise performance tests [VSE 8.7i P2 vs P3 testing]
                      brentil

                      Thanks for the compliments.  This just seemed like an important bit of information that was lacking for this product set to the public.  Like I said, feel free to suggest benchmark ideas for future tests.

                      • 8. Re: VirusScan Enterprise performance tests
                        brentil

                        Rebuilding my testing system to be a bit more modern, Core 2 Duo and 4GB of RAM.  I will be doing 8.7i P3 versus P4 testing as soon as I have that done.  I may also throw in new command line comparisons as well since I believe there is a new version of that as well.

                        • 9. Re: VirusScan Enterprise performance tests
                          brentil

                          New machine with new tests.  I've added two Java based tests as well since these are the most impacted by Anti-Virus applications.

                           

                          Machine Configuration was  as follows;

                           

                          Dell OptiPlex 745, Core 2 Duo 2ghz, 4GB of RAM
                          W7 RTM 64bit fully patched
                          VSE 8.7 + AntiSpyware module
                          Office 2010 fully patched, VS2010  fully patched, SQL2008 R2 Management Suite fully patched
                          FireFox, Chrome, Flash, Java 1.6, TortoiseSVN, Eclipse, Oracle SQL Develope

                           

                          I've created a copy of the Program Files & Program Files (x86) folders post installation resulting in ~42,000 files and 7 GB of data.  This data is used for the tests done.  Each test is run 3 times and averaged together.

                           

                          Copy - An xcopy command is used to make a complete duplicate of the files into another folder

                          RAR - The command line WinRAR application is used to make a default compression RAR file of the files

                          unRAR - The command line WinRAR is used to decompress the RAR file

                          Eclipse - The Windows Eclipse executable is loaded and timed via stop watch

                          Oracle - The Windows Oracle SQL Developer executable is loaded and timed via stop watch

                           

                           

                          on 1/6/11 11:41:41 AM GMT-05:00
                          1 2 Previous Next