Layer 1 / Layer 2 Difference

T1 is Layer-1. It describes the physical electrical characteristics, as well as the framing and data encoding format.

Frame Relay, HDLC and PPP implement Layer-2 protocols over the T1.

ATM is an odd one, it doesn't quite fit things, but layer-2 is the best place for it, although there are some layer-1 ATM bits.

The service provider doesn't do all that much with T1s, its more along the lines of pass the bits, and supply repeaters that look for something like a T1 and pass it on. Both customer ends do more to actually specifying what a T1 is.

Hi Doug,

thanks for the reply. I am trying to understand that when service providers provide a Layer 1 Circuit for eg (T1, T3, OC-3 etc) than who provides the layer 2 in that case.

When the service provider gives a Frame relay connection they are providing services upto Layer 2. But in the case of Layer 1 when they provide a physical circuit who drives the layer 2 portion.

The only thing i can think of is the customer either specifying PPP or HDLC on the router for encapsulation. Is that what would constitute as layer 2 when the service provider just provides a physical circuit.

It's all relative.

The layered network models are tools that need to be customised and applied to your needs.

The easiest cases are, typically for us, IP over Ethernet. In that case you know that the cable and the data signalling are at layer 1 etc.

Things get confusing as more sophisticated systems are considered. For example ISDN has a phy, bit signalling, encoding, addressing, routing ... if you do a show "isdn status" is says "xx layer 3 calls active" A typical IP engineer though sees ISDN as a Layer 1 technoloy (well I do anyway) which he then uses to send his bits over.

Here is an analogy I just made up. In the air transport business (well this one anyway)

Carrier Pigeons are placed in Layer 1 of the Open Air Transport (OAT) Reference Model. Crated goods are in Layer 2 A managed end to end delivery service in Layer 3.

Jim calls up a carrier pigeon to deliver a letter and a few days later gets confirmation that his letter has arrived.

What Jim doesn't know and indeed //doesn't care about// is that Ms Pidgeon used her Gold amex card and travelled in a luxury box on a private jet to the destination spending the balance of the time resting up on the beach eyeing up the peacocks.

All Jim //does// care about is that his Pigeon interface works as expected and that the information arrives.

MPLS for examle can be used to provide an apparent Layer 1 service (i.e. it looks EXACTLY like a wire to the two end parties if they do not look too closely) but in reality the data is carried over a routed IP network. [Looking closely would involve for instance sending in bad Ethernet frames (L2) and observing that perhaps they did not come out the other end. If it was really a wire the bad frames would come out the other end.]

I am on shakier ground here but here goes:- Similarly a working ISDN interface can be considered like a layer 1 bit carrier for PPP. When the ISDN is broken and you want to fix it, then a different approach may be helpful. You may need to know that ISDN has L3 addressing (phone numbers) etc.

ATM is a fully featured network with addressing, resilient routing, and many other features however as an IP engineer ATM is largely used to simulate a wire, which is indeed one of its design goals.

The absolute key thing is that there are layers and that the layers operate independently (mostly/ideally:-) As an IP designer/troubleshooter this independence can facilitate simpler and more managable network designs and easier behaviour isolation.

Finally - Jim's Boss may care about the details of the service implementation since such details can affect the reliability of the service. We are back again to that flexibility thing. Apply the //model you need// to //get your job// done.

I am sure that this is very confusing but I hope that I have helped a little. Stick with it, I find the Layered model very useful in my work. I find it assists very much in creating the right questions to ask. Then it is only a matter of determining the answers.

-- Me -- CCEI #TBA I hope no one minds my joke here. I have just begun on the trail and hope that reminding myself of it here will help keep the nose to the grindstone.

Depends on who does what, and what kind of service.

The telco really hands off a connection at two sites, and repeats a signal that looks like its structured like the one you've purchased from them with enough bandwidth to handle what you've paid.

The Layer-2 really starts with your router one one side going to the remote site, and then the remote site back to you.

The telco typically will have two groups. One that runs the WAN switch, and one that provides the pipe. You could consider the group that runs the WAN switch to be providing the layer-2, but really its just a piece of gear like you have, but its centeralized to probably save you money by combining it all done to your connection.

Both sides "drive" the layer-2 portion.

Yes, the service provider configures their side to match the requirements, and you setup your side to match the requirements.

Ethernet protocol is layer-2. The fiber would be layer-1 in MOE. Again both sides really "drive" it. Typically the network provider will implement ethernet switches at both ends that they control to monitor/shape your access into the MOE network.

Depends on the type of service. Point to point circuits need layer 2 provided by the customer. In the case of frame relay, both the customer and the telco provide layer 2 services.

As mentioned above, layer 2 is provided by both the customer and the telco. The customer will speak layer 2 (lmi) with the telco at each location. Then the telco creates virtual circuits to connect each site as needed.

Yep.

A metro ethernet service does offer layer 2 from the telco. These services are atm and mpls based within the telco network and are passed to the customer via ethernet.

Jim