[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
WGLC ent-analysis-03: application vs network driven
(I'll put two most separable & possibly most discussable items in separate
mails.)
==> the document seems to have been written based on the assumption
that first the enterprise decides which kind of network it will run
(dual-IP, v6-only) and whether its applications will be v6-only,
dual-IP or v4, and then takes a look at the surrounding environment --
and the result is often "oops.. we need translation, tunneling, and a
number of other complexities for our decision to be fulfilled".
In some cases, this approach may be appropriate (though I hope this is
not commonplace). In most cases, the better way is to look at what
kind of applications you have or want to have (also the ones which you
need to communicate with outside the enterprise) and after that make
the decisions (weighing the different tradeoffs of course) which kind
of network topologies would be best suited for the job.
After this deliberation, the result might be that an enterprise might
have to take a "hard road" in any case, resulting the matrix we're
describing, but the thought process is different.
Maybe or maybe not this has also been implicit when writing this
document, but this thought process just hasn't been spelled out. It
should be, in a number of places. I think a key message that needs to
be delivered is, "if a particular scenario is difficult, maybe it is
worth considering alternative assumptions?" [for example, having the
application be dual-IP or network dual-stack] or in other words,
"don't push if there's resistance".
This issue should probably be discussed in section 3, maybe reminded
briefly in section 6, and again mentioned in appropriate places in
section 8.
--
Pekka Savola "You each name yourselves king, yet the
Netcore Oy kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings