none
WHCK setup for Windows 8.1 RRS feed

  • Question

  • Hi,

    I have previously tested on windows 8 and 8.1 preview, recently i have setup for windows 8.1 pro w, there we used external AP's ( 2 RT-N66U and 1 RT-AC56U ) and LAN of those AP's( RT-N66U recommended by microsoft) are connected.

    I Have some doubts on WHCK setup

    1> Why external Ap's are used

    2> Why LAN connection is used in between two AP's

    3>why we require Two Ethernet adapters in AP controller side

    4> Why we require two Gigabytes supported switches with different IP Address

    5> I passed 165 Test cases out of 168, only three  WIFI Direct performance related test cases are failing, how to debug these test cases?

    Thursday, January 30, 2014 2:04 PM

Answers

  • Hi, I will reply to each of the questions you asked below. I include a contact address below for any further questions also.

    1> Why external Ap's are used

     - Because we have found more bugs in wifi miniports by testing against real APs that can be purchased on the market, thus, somewhat better interoperability is tested (previous versions of our tests used artificial APs based on special network adapters for almost all test cases)
     - These APs are able to perform many more functions than any virtual AP might be able to and thus cover more user scenarios in the HCK
     - Our ability to remotely configure the APs allows for easier trasition between the test scenarios


    2> Why LAN connection is used in between two AP's

     - Some of the test scenarios involve simulating an enterprise roaming scenario (a user walking down the hall at work), our tests need to ensure both APs have a hard connection between them when testing the scenario since the TCP session is tested across the APs. The test code will ensure that the connected ports are disabled during all other tests. During the tests which use the connection, DHCP is disabled on one of the APs as well. This closely approximates an actual enterprise deployment.

    3>why we require Two Ethernet adapters in AP controller side

     - The primary reason for this is that the performance tests (wlan devices have very specific throughput requirements they need to meet) should have their own dedicated network so that there is no other background traffic interferring with the measurements.
     - However, secondary to this, because of how the APs are controlled remotely, they need their own address space seperate from whatever network the SUT/DUT/HCK are on.
     - As mentioned in the documentation, some of the tests require that the HCK network be assigned by DHCP in a certain range as well. This is so the test can automatically determine which network adapter is used for which function.

    4> Why we require two Gigabytes supported switches with different IP Address

     - The primary reason for this was to keep the equipment ordering process simple.
     - Technically, only the switch connecting the APs must be gigabit, the reason being that some of our performance requirements test speeds faster than 100Mbps, which means that a 100Mbps switch would fail the performance test.
     - But keeping them both the same keeps setup and purchasing much simpler.

     - The reason that the address range MUST be seperate is that all computers when connected to two networks must have non-overlapping IP address ranges. This is just a requirement of how TCPIP works, so that the computer knows which adapter to send the packet (or response on).
     - In addition, due to the configuration of the APs, the AP side of the AP machine must be configured statically (no DHCP), and the HCK side of the AP machine must be configured with DHCP, again the reason being so that our test can determine which adapter is connected to which network, since our tests need to send data on each seperately.

    5> I passed 165 Test cases out of 168, only three  WIFI Direct performance related test cases are failing, how to debug these test cases?

     - The quick answer to this, is that information about WHY the tests fail is always contained in the test logs.
     - Test logs for each test can be accessed by opening their results on the results tab and in most cases, right-clicking on each result to open the log files.

     - For the Wifi-Direct performance tests, the best advice is to ensure enough performance to consistently pass the tests. These three tests specifically will test several scenarios in each of the three tests which all must meet performance requirements.

     - There may be a way we can help further though, but we would need to confirm first that the three tests you mention are indeed passing sufficiently. Please send an HCKX file with the three failures to wlanndt __AT__ microsoft.com for review and we can take a look further. (or in the case of a large file, uploading to onedrive or skydrive works well too, but simply share out a link.)


    - Any further questions can also be directed at the above address as well. (We publish the same email address in our documentation as well if you want to confirm, which you can find here:)

    http://msdn.microsoft.com/en-us/library/windows/hardware/jj123847.aspx

    J.P.

    Friday, January 31, 2014 7:59 PM

All replies

  • Hi, I will reply to each of the questions you asked below. I include a contact address below for any further questions also.

    1> Why external Ap's are used

     - Because we have found more bugs in wifi miniports by testing against real APs that can be purchased on the market, thus, somewhat better interoperability is tested (previous versions of our tests used artificial APs based on special network adapters for almost all test cases)
     - These APs are able to perform many more functions than any virtual AP might be able to and thus cover more user scenarios in the HCK
     - Our ability to remotely configure the APs allows for easier trasition between the test scenarios


    2> Why LAN connection is used in between two AP's

     - Some of the test scenarios involve simulating an enterprise roaming scenario (a user walking down the hall at work), our tests need to ensure both APs have a hard connection between them when testing the scenario since the TCP session is tested across the APs. The test code will ensure that the connected ports are disabled during all other tests. During the tests which use the connection, DHCP is disabled on one of the APs as well. This closely approximates an actual enterprise deployment.

    3>why we require Two Ethernet adapters in AP controller side

     - The primary reason for this is that the performance tests (wlan devices have very specific throughput requirements they need to meet) should have their own dedicated network so that there is no other background traffic interferring with the measurements.
     - However, secondary to this, because of how the APs are controlled remotely, they need their own address space seperate from whatever network the SUT/DUT/HCK are on.
     - As mentioned in the documentation, some of the tests require that the HCK network be assigned by DHCP in a certain range as well. This is so the test can automatically determine which network adapter is used for which function.

    4> Why we require two Gigabytes supported switches with different IP Address

     - The primary reason for this was to keep the equipment ordering process simple.
     - Technically, only the switch connecting the APs must be gigabit, the reason being that some of our performance requirements test speeds faster than 100Mbps, which means that a 100Mbps switch would fail the performance test.
     - But keeping them both the same keeps setup and purchasing much simpler.

     - The reason that the address range MUST be seperate is that all computers when connected to two networks must have non-overlapping IP address ranges. This is just a requirement of how TCPIP works, so that the computer knows which adapter to send the packet (or response on).
     - In addition, due to the configuration of the APs, the AP side of the AP machine must be configured statically (no DHCP), and the HCK side of the AP machine must be configured with DHCP, again the reason being so that our test can determine which adapter is connected to which network, since our tests need to send data on each seperately.

    5> I passed 165 Test cases out of 168, only three  WIFI Direct performance related test cases are failing, how to debug these test cases?

     - The quick answer to this, is that information about WHY the tests fail is always contained in the test logs.
     - Test logs for each test can be accessed by opening their results on the results tab and in most cases, right-clicking on each result to open the log files.

     - For the Wifi-Direct performance tests, the best advice is to ensure enough performance to consistently pass the tests. These three tests specifically will test several scenarios in each of the three tests which all must meet performance requirements.

     - There may be a way we can help further though, but we would need to confirm first that the three tests you mention are indeed passing sufficiently. Please send an HCKX file with the three failures to wlanndt __AT__ microsoft.com for review and we can take a look further. (or in the case of a large file, uploading to onedrive or skydrive works well too, but simply share out a link.)


    - Any further questions can also be directed at the above address as well. (We publish the same email address in our documentation as well if you want to confirm, which you can find here:)

    http://msdn.microsoft.com/en-us/library/windows/hardware/jj123847.aspx

    J.P.

    Friday, January 31, 2014 7:59 PM
  • Thank you J.P

    In some test cases external AP's will be used ,How these AP's will be configured (through AP controller or DUT )?

    I merged two packages.In newly created package, number of test cases are doubled. In selection tab it is displaying same device in two times with different group. why test cases are doubled in created package?

    Can we test manually all wifi direct test cases?

    Regards

    Dhananjayakumar

    Tuesday, February 4, 2014 7:14 AM