Major <Major.30rbqe@no-mx.wirelessforums.org> hath wroth:
>Hi Everyone!
On behalf of everyone, welcome.
>I have recently decided that there is much gold/fun/travelling to be
>had in the wireless netwroking department.
Gold? Yes, but only to those that are at the top of the profession.
That days when wireless was magic and required high paid magicians are
over.
Fun? Well, I have a rather warped idea of what constitutes fun. If
staying up all night troubleshooting defective radio or protocol
implementations is your idea of fun, then by all means, have fun.
Travel? Well, wireless tends to be the "solution" to many 3rd world
communications problems. The problem is not the lack of
infrastructure. It's that any infrastructure that can be sold, is
immediately stolen. For example, copper wire disappears as soon as
it's installed. Fiber isn't much better as it also has value. That
leaves wireless, were only the end points need to be guarded. Yeah, I
think you'll just love 3rd world wireless.
>primarily I am concerned with
>wireless protocols :
>3gpp, umts, Wcdma etc.(as well as the upcoming 802.11n)
Take a giant step backwards. Wireless comes in 3 parts, all of which
you should obtain some experience in handling. There's RF (radio),
protocols (networking), and installation (towers, topography, and
propagation). If you concentrate on just the protocol aspects, you've
already over specialized. I've lost count of how many users in this
group are stuck on troubleshooting a "packet loss" or "disconnect"
problem at the operating system or networking level, when the problem
is really in the antennas, propagation (through walls), or
installation. Oh yeah, some experience in interference effects, test
equipment operation, stealing coax adapters, and diagnostic tools are
also very useful.
>and due to the
>already exapnding options I was wondering if anyone knew where would be
>a good place to start. I have limited wireless knowledge but a thirst to
>learn it all.
Thirst can be satisfied one glass at a time. If you try to guzzle the
entire subject at once, you'll drown, or at least end up with a
serious hangover. Take it small step.
>Does anyone have any pointers of where to start (I am at uni and so
>have about 3 years worth of time I can devote to this)
Learn by Destroying(tm). I mean that. All you're ever going to have
time to do at skool are the fundamentals. You'll probably get very
little practical experience. If possible, sell your spare time to the
local WISP (wireless internet service provider) at slave labor rates
in order to gain experience. You'll learn quite a bit including much
that you can apply from what theory you've learned in skool. At
skool, do something practical with wireless. Setup your own wireless
network, with real live complaining students as customers. (I did
this is college when we wired the dormitories for telephone and
installed a noisy PABX switch). Don't worry about what to learn. That
will be very apparent as soon as things start to fail or don't work
the way you expect.
In class, I suggest a program in math, computers, and simulations.
You'll find that these form the basis of most design work. The days
of cut-n-try are largely over. Get some experience building an RF
PCB. It can be anything at microwave frequencies. It probably won't
work the way you designed it. The learning is in the troubleshooting
and re-design, not in the initial design.
Once you have the math moderately in hand, work on basic circuit
design. It doesn't have to be anything fancy, just learn what the
components do, their limitations, where to buy them, how to beg for
samples, and how to read spec sheets full of lies and distortions.
Incidentally, the data sheets and the real components specs are
usually close, but no more.
At the same time, you're going to need some experience with networking
protocols. Don't get hung up learning every little detail about
TCP/IP. Try to get a general feel for timing issues, performance
determining issues, oddities like satellite delays, and the effects of
transmission impairments. If you run into some really strange
non-obvious protocol design feature, that's worth investigating.
Also, you'll need to read the FCC Part 15 (especially 15.247) rules,
and the various IEEE 802.11 specs. Note that I said read, not
understand, as I don't thing anyone alive really understands the legal
manure that masquerades as technical specifications. If you plan on
actually producing a real product, that must meet ISO, FCC, IEEE, etc
specs, you'll need to read and perhaps understand this stuff. Save it
for later as it tends to be mind numbing.
Frankly, I don't think you can do it in 3 years of study from scratch.
Methinks 4 would be about right. Best of luck.
>any documents
>I should concentrate on or books that would help out. My aim is to
>eventually write protocols myself - so if anyone does this themselves
>could you tell me how you got into it, where you started and possible
>links. Thanks for all and any help
There's an old adage about not wanting to witness how sausage and laws
are made. The same applies to standards.
Once upon at time, in a land far far away, I helped scribble what was
to be a proposed EIA standard. Suffice to say that the real standard
was hashed out at company expense, inscribed and edited by copany
attorneys, and made little technical sense when it was submitted for
the initial round of hysterical verbal abuse. Fortunately, the
lawyers were all too drunk during the meeting to present much of a
defense, so us techys locked ourselves in a hotel room and hashed out
a workable compromise in an all night marathon. The final proposal
was submitted by those that remained sober and standing.
These days, it gets really messy and ugly. One look at the history of
802.11n protocol standards should give you a clue that it's really an
exercise in politics, patent law, and litigation, not technology. For
example, 802.11n contains two completely different technologies, that
should never have been in one standard, except that both proponents
needed the votes from the other technology mob to get the proposal
passed.
Careful what you wish for. It may be your next nightmare.
Anyway, one of the few really good things that instructors can offer
is a recommended course of study. They have books, experience, and
usually a clue as to what is required and expected. I suggest you ask
the same question to a qualified instructor. If nothing else, you may
get to borrow the books for free (I did) instead of having to buy
them.
--
Jeff Liebermann
jeffl@cruzio.com
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060
http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
Re: Budding wifi protocol'er 
Linear Mode
