Traditional classful IP addressing?

ocicat002 · ‎11-14-2007

In classful addressing, class A networks would allow up to 16,777,214 hosts be attached to the network. In reality, how was this managed? This many hosts were attached to a single LAN? Was the traffic so light that this many hosts could be attached? Were there no routers used to break up collision domains? As I understand, any bridging employed would only compartmentalize broadcasts.

Any insight which can be shared would be appreciated as this seems to be an unreal number of hosts to have on a single segment.

keller.oliver · ‎11-14-2007

Although I wasn't there back in the time :), I guess they did a lot of subnetting.

One of the earlier RFCs talks about a Class A like structure, where you had fixed structure of the IP address : 8 bit network, 24 bit hosts. My suspicicion is that in the early days of ARPANET, they just thought that there never would be more than about 100 networks :D. I don't think they expected 16 million hosts in a single segment.

Quote from that historical RFC 760 :

"Addresses are fixed length of four octets (32 bits). An address begins with a one octet network number, followed by a three octet local address. This three octet field is called the "rest" field." (page 7)

Best wishes,

Oliver

ocicat002 · ‎11-15-2007

What I haven't been able to confirm is *when* subnetting was made available. Was it available at the advent of ARPANET & beginning of IP networks? If so, then managing Class A networks makes sense with subnets. If some time lapsed between the beginning & the introduction of subnets, I don't see how traffic over large networks could be controlled.

Any clarification anyone can provide would be appreciated. Thanks.

dinesh_mih · ‎11-15-2007

Hi,

I think at that point of time since a lot of IP addresses were available, people were easily assigned class A IP addresss so for example a company is assigned 11.0.0.0 IP address then 11.0.0.1 till 11.255.255.255 IP belongs to him. I am sure noone would have connected so many IP addresses in a single LAN. The inside LAN will be a big structured network with lot of broadcast domains by using routers and layer 3 devices. It is just that for the outside world (internet) the full 11.0.0.0/8 range belongs to one compnay. Ofcourse inside that company the IP addresses were sumnetted.

Regards

Dinesh

http://Knowurtech.com

(Good technical articles. Submit your articles and get featured on our website. Send email to articles@knowurtech.com. To know the benefits http://www.knowurtech.com/be_an_expert.html)

Hi,

I think at that point of time since a lot of IP addresses were available, people were easily assigned class A IP addresss so for example a company is assigned 11.0.0.0 IP address then 11.0.0.1 till 11.255.255.255 IP belongs to him. I am sure noone would have connected so many IP addresses in a single LAN. The inside LAN will be a big structured network with lot of broadcast domains by using routers and layer 3 devices. It is just that for the outside world (internet) the full 11.0.0.0/8 range belongs to one compnay. Ofcourse inside that company the IP addresses were sumnetted.

Regards

Dinesh

http://Knowurtech.com

(Good technical articles. Submit your articles and get featured on our website. Send email to articles@knowurtech.com. To know the benefits http://www.knowurtech.com/be_an_expert.html)

andrew.burns · ‎11-15-2007

Hi,

Subnetting became "officially" available with the publication of RFC 917 in Oct 1984 but the practise was in use before then. I think initially, before anyone thought of subnets, the practise was to assign host numbers without regard to which LAN a host was on. (Although, I'm not entirely sure how that worked...) And don't forget, Cisco was only founded in 1984.

HTH - plz rate if useful.

Andrew.