none
determine if nat64 is used or native ipv6/isatap RRS feed

  • Question

  • Hello there,

    as my previous benchmarks with teredo or ip-https tunnels were not that great, speaking of network performance i was told by Yaniv that the nat64 protocol has a huge overhead. well, i thought about that and wanted to test it, but without enabling isatap in my whole domain. So I enabled isatap just on one vista x64 machine by adding the isatap entry to the local hosts file. rebootet and voila, virtual isatap interface came up online. So i connected again from outside to this special isatap-enabled machine but encountered no change in network performance. so my question would be: how can i determine for sure if nat64 protocol is used or native ipv6/isatap? How can i find this out?

    best regards,
    Jörg
    Tuesday, January 12, 2010 9:46 AM

Answers

  • Hi Joerg,

    The overhead is in CPU on the UAG server.
    There will probably be no affect on network performance if you use NAT64 or ISATAP.

    But using ISATAP, the UAG server's CPU is used much more efficiently and is able to support a larger number of clients.
    • Marked as answer by Erez Benari Tuesday, January 12, 2010 8:06 PM
    Tuesday, January 12, 2010 2:01 PM

All replies

  • Hi Joerg,

    The overhead is in CPU on the UAG server.
    There will probably be no affect on network performance if you use NAT64 or ISATAP.

    But using ISATAP, the UAG server's CPU is used much more efficiently and is able to support a larger number of clients.
    • Marked as answer by Erez Benari Tuesday, January 12, 2010 8:06 PM
    Tuesday, January 12, 2010 2:01 PM
  • hmmm...that means directaccess via teredo or httpstunnel is definitely about 40% slower than a standard pptp connection. this is not that great ;-( - do you have any other possibilities or ideas to optimize/speedup the performance?

    best regards,
    Jörg
    Tuesday, January 12, 2010 8:08 PM
  • come on guys, any idea? i can hardly explain the great features of directaccess (which is great in my opinion, very smart, very easy and for the enduser very easy, too) to anyone when they realize it´s about 40% slower than pptp. i am sure there are optimization possibilities. anyone?
    Wednesday, January 13, 2010 3:04 PM
  • On what are you basing the claim that it's 40% slower?

    Wednesday, January 13, 2010 4:12 PM
  • i benchmarked it. i connected via pptp to my internal network and did a smb transfer. after that i connected via directaccess, first via teredo, then via httpstunnel, did the math and discovered an about 40% decrease in the teredo and the httpstunnel compared to the pptp connection.

    best regards
    Joerg
    Thursday, January 14, 2010 12:04 PM