Re: TCP end-to-end Semantics

From: Craig Partridge ([email protected])
Date: Wed Jan 10 2001 - 09:58:56 EST

  • Next message: Manish Karir: "Re: TCP end-to-end Semantics"

    In message <[email protected]>, Manish
    Karir writes:

    >well I believe the question is more along the lines of: if ISP X uses
    >spoofer or caches or whatever to give you better performance(thruput or
    >security or whatever you are measuring performance as) then you will
    >continue to use them. If their service to you is unacceptable then
    >you as the user can choose to switch to ISP Y. The bottom line is
    >will the free market support for intelligence in the middle applications
    >be sufficient for them to survive...
    >if ISP X notices 10 big customers switching to someone else believe me he
    >will quickly drop spoofing.

    This assumes that the customer understand the source of error.

    One of the issues that has cropped up is that customers have expectations
    of their ISP. The introduction of hidden services like spoofers that
    may not conform with customer expectations but are not made visible to
    the customer means the customer may not get the service he/she expects but
    not know the cause. (E.g., the customer may switch and get another ISP
    that does the same hidden thing, and see the same problems, and blame the
    technology not the ISPs).

    Craig



    This archive was generated by hypermail 2b29 : Wed Jan 10 2001 - 10:34:20 EST