Re: Performance strategies
From: Ivan Medvedev [MS] (ivanmed_at_online.microsoft.com)
Date: 07/07/03
- Next message: Ivan Medvedev [MS]: "Re: Checking for SecurityAttributes"
- Previous message: Ivan Medvedev [MS]: "Re: Decryption Error: PKCS7 padding is invalid and cannot be removed"
- In reply to: Doman Maciejko: "Performance strategies"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Mon, 7 Jul 2003 14:49:53 -0700
Doman -
when running managed applications in a fully trusted environment (such as a
local drive) the performance overhead caused by security is very minimal and
mostly condenced in the startup time. I doubt game developers will be
turning the security off because it does not buy them much and because of
the publicity such games will have (nobody will like to buy a game that
compromises system's security). Also, not providing APIs for
changing/switching off the security would impose unnecessary hardship on
those writing legitimate administration tools while not preventing arbitrary
software running fully trusted from doing that (because almost everything
that can be done by a computer operator can be as well done by software).
--Ivan
This posting is provided "AS IS" with no warranties, and confers no rights.
"Doman Maciejko" <doman.is@home.se> wrote in message
news:OKRhhfTQDHA.2432@TK2MSFTNGP10.phx.gbl...
> It is recommended to designing code to perform as few stack walks as
> possible to gain performance. But is there any statistics which show the
> phenomena?
>
>
>
> An alternative, effective but extremely dangerous and not recommended way
to
> gain performance is to turn off the security manager. I assume that
> Microsoft has carefully considered how such a high risks operations could
be
> misused before implementing them. But how much performance is it possible
to
> gain by turning the security manager off?
>
>
>
> An obvious risk area is game developing where CPU-cycles are hard
currency.
> Game developers will probably deliberately turn off the security manager
to
> gain performance, is that considered? Does Microsoft presume that
developers
> will turn off the security manager in general, mainly to gain performance?
>
>
>
> Wouldn't it be safer in a way just allowing the administrator to turn of
the
> security manager and not letting it be done programmatically? Which
> complications could such as strategy bring?
>
>
>
> I hope I ask these questions in the right forum, otherwise I would gladly
> accept hints about where I could find answers to my questions :-)
>
>
>
> /Doman Maciejko
>
>
>
>
- Next message: Ivan Medvedev [MS]: "Re: Checking for SecurityAttributes"
- Previous message: Ivan Medvedev [MS]: "Re: Decryption Error: PKCS7 padding is invalid and cannot be removed"
- In reply to: Doman Maciejko: "Performance strategies"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|