RE: Time to kill layer 2

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



The BOFH way - kill the user and let the network be.... ;)

 

Kidding of course

 

 


From: Carlos M. Martinez
Sent: 15 April 2016 08:18
To: Ted Lemon;Dimitri Staessens
Cc: ietf
Subject: Re: Time to kill layer 2

 

 

All the way up to the user ? Many times.

 

On 4/14/16 6:56 PM, Ted Lemon wrote:

> Of course!

>

> On Thu, Apr 14, 2016 at 5:54 PM, Dimitri Staessens

> <dimitri.staessens@xxxxxxxxxxxxxx

> <mailto:dimitri.staessens@xxxxxxxxxxxxxx>> wrote:

>

>     anyone thought of killing everything on top of layer 2?

>

>

>     On 04/14/16 14:59, Phillip Hallam-Baker wrote:

>

>         This morning I spent an hour debugging the network to print out two

>         class projects that were due. Some points:

>

>         1) My ability to debug the network is better than 99% of the

>         population

>         2) The interaction of Bonjour, DHCP and auto power saving is

>         unfortunate

>         3) Things should still work after I have been away for a week

>         4) If vendors want to be selling all that IoT gear, they have to

>         solve

>         these issues.

>

>         5) I want someone to blame. Right now when the network doesn't

>         work, I

>         don't know who is the cause. I want one point of contact. Whoever is

>         that point of contact will get most of my networking money.

>

>

>         One of the biggest headaches in debugging is that 'smart hubs' are

>         not. They are actually very stupid. They make assumptions of network

>         topology that are not true. Another is the unfortunate

>         implementation

>         of DHCP.

>

>         I don't use SNMP for a simple reason - it is not available to most

>         ordinary people. I want to understand networking for the 99%,

>         not the

>         IETF 1%-ers.

>

>         All this networking gear is presented to me as black boxes over

>         which

>         I have absolutely no control (which is fine-ish) and no visibility.

>

>         What we have today is the product of a historical process. I

>         remember

>         the days when Ethernet ran on 10BaseT. But I installed my first

>         switch

>         30 years ago and it has been a switched protocol for 20 years now.

>

>

>         It seems to me that there is a business opportunity for any

>         vendor who

>         takes the rather obvious step of simplifying the system.

>

>         People talk about 'IP everywhere' and 'IP end-to-end' which is

>         rather

>         odd when you think about the fact that virtually every local network

>         uses MAC addresses for routing.

>

>         One of the reasons that IP won against OSI was that it was simpler.

>         Applications ran on top of the IP layer with only TCP inbetween. Of

>         course these days we do have a Presentation layer, Web Services

>         run on

>         HTTP. But unlike the OSI presentation layer, ours does not introduce

>         extra moving parts.

>

>         It seems to me that if we really believed in IP everywhere and IP

>         end-to-end we would insist that network switches be IP routers that

>         can be managed using BGP/OSPF or at least routing tables rather than

>         heuristic devices that try to guess where packets should go based on

>         goat entrails, phases of the moon or whatever they use.

>

>

>         What should have happened many moons ago was that DHCP should have

>         become a bidirectional protocol or a bootstrap to a bidirectional

>         protocol. So when a printer joins the network, it authenticates and

>         tells the network what it is. And this is all defined in one set of

>         specifications from one organization, none of which assumes that

>         security is an 'advanced', 'optional' or 'enterprise' feature.

>

>         Instead we have an ad-hoc layer trying to achieve the same result in

>         peer-to-peer fashion. A similar approach works for frogs as a

>         reproductive mechanism but only at the species level. It certainly

>         does not work for the individual ova which may or may not connect to

>         the printer it is trying to use to print the kids damned homework.

>

>

>         Seriously, the fact that things have scaled thus far and the 1% can

>         get them to work does not mean that we can get to the next level

>         without a serious rethink of the local network architecture.

>

>         The type of device I think we need would be first and foremost an IP

>         router. It would have ethernet plugs on the box and use ethernet

>         layer

>         1 specs. But when a another 'True-IP' device was plugged in, it

>         would

>         quickly negotiate a direct IP connection, oh and with proper 64KB

>         packets. It would also, authenticate, announce and turn on link

>         layer

>         encryption.

>

>         Such a device would also be a legacy router. It would fake all the

>         signals necessary for a legacy ethernet device to function. It would

>         also be responsible for maintaining the local information for the

>         network service database and intercommunicating with other hubs to

>         achieve a global network view.

>

>

>         The net result of all this would be that I would never ever need to

>         install another printer (no, it is not actually necessary for every

>         stupid printer to have its own stupid printer driver). Opening the

>         'printers' folder would automatically show every printer that is on

>         the network or can be woken from slumber by the hub it connects to.

>

>

>

 

 

 


[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]