D&C GLug - Home Page

[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]

Re: [LUG] OT: Fwd: Net approaches addressing limit

 

On Thu, May 13, 2010 at 8:58 AM, Gordon Henderson
<gordon+dcglug@xxxxxxxxxx> wrote:
> On Wed, 12 May 2010, Julian Hall wrote:
>
>> On 12/05/2010 16:41, Grant Sewell wrote:
>>>
>>> Token Ring is also a dead technology, but it is still taught so that
>>> students have a grounding in *concepts* other than CSMA/CD.
>>>
>>> Just because something is dead doesn't mean it is not *useful* to learn
>>> about it - either from a historical point of view, to introduce
>>> concepts or to stagger the learning process.
>>>
>>> I still teach classful addressing for IPv4 - Class A, B, C, D&  E.  It
>>> would be kinda hard to explain why the multicast address ranges begin
>>> with 224 and are called "Class D" addresses without visiting classful
>>> addressing.  A little way down the line I introduce CIDR and related
>>> topics and not one student has become confused by it.
>>>
>>> Grant.
>>>
>> Agreed.  I didn't take the Networking stream in my degree, but the
>> Information Processing and Management module did cover some of the basic
>> network topologies including Token Ring, Star etc.  They may have taught
>> Class addressing in the full Network stream and I agree that concepts are
>> just as important to learn so as a graduate you don't think 'OK I know where
>> we are, but how and why did we get to this point?'
>
> I'm still going to disagree. Classed addressing is dead, gone and buried.
> It's not even pining for it's Fjords - it's a concept no-more and no-one is
> missing it.
>
> Seriously - no-one in todays world uses classed IP addressing. You have a
> network and a subnet mask or /prefix and that's it. That covers the whole
> range. You do not have to remember anthing special at all. If you (Grant)
> actually get students to build a router with a Class C on one side and a
> Class A on the other and make them route then you're wasting your time and
> stuck in an age that no-one in the real world is intersted in.
>
> Todays Internet use CIDR in both the IPv4 and IPv6 worlds.
>r
> Classed addressing is history, and it's 17 years of history too - CIDR was
> introduced in 1993 and by 1996 all the ISPs I was working with were refering
> to it rather than the old classes.
>
> Seventeen years! That's a lifetime in modern Internet terms. There are
> people on this list who were not alive when CIDR was developed - why burden
> them with useless history.
>
> Teaching it as something that's actually used, and using it as a precursor
> to teaching classless IP addressing is a total waste of time.
>
> While I am stuck in my ways on some things, old time classed IP addressing
> is not one of them. Get with the times - 17 years ago!
>
> As for token ring - it's the Betamax of the networking world and it would
> have taken off if the vendors hadn't been so snotty about it and charged an
> absolute fortune for the kit. (And FWIW; I was using token ring kit with PCs
> and Transputer based control boards in 1989) That deserved to die, not
> because the networking technology was inferior (it wasn't), but because
> mechnically it was a shambles.
>
> Gordon

I think we're going to have to agree to disagree on this.  You have
some very valid points, but I am still of the steadfast opinion that
simply abandoning a useful conceptual tool merely because a more
complex concept has superceded it is daft.  I do agree that teaching
people classful addressing and routing and leaving it at that won't
necessarily help them in the slightest in today's classless world...
but you don't teach people to swim by asking them to cross the channel
on day 1.  If nothing else, teaching people about classful addressing
will stop the questions of "why's it called *classless* inter domain
routing?".

Grant.

-- 
The Mailing List for the Devon & Cornwall LUG
http://mailman.dclug.org.uk/listinfo/list
FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html