1

Studying the WAN lines types, am I right regarding the statements below, or if not, could you correct me please:

  • A real leased line, could be performed with TDM on a line with for example SDH, TDH or SONET with a CSU/DSU on the customer site (a layer 1 leased line?)
  • A leased circuit could be done with a ATM CPE on the customer site, typically for a point to point connection needs (a layer 2 circuit?)
  • A leased circuit could be done with a MPLS tunnel with what we call Provider router, and Customer Router on the customer site, for any to any connection type

On the case of ATM and MPLS, as far I understood it's like "logical" leased line, and how the ISP reserve the bandwidth for the customer ? And what could be the layer 1 protocol ? SDH, SONET ? Other ?

And finally, generally, what is the most frequent technics implemented, leased line or leased circuit ?

Many thanks, all these notions are not very clear for me.

4

A real leased line, could be performed with TDM on a line with for example SDH, TDH or SONET with a CSU/DSU on the customer site (a layer 1 leased line?)

This depends - usually a (rare) leased physical line is just copper or (dark) fiber. It's up to you what you do with it.

A leased circuit could be done with a ATM CPE on the customer site, typically for a point to point connection needs (a layer 2 circuit?)

Yes, for instance. ATM is pretty dead though.

A leased circuit could be done with a MPLS tunnel with what we call Provider router, and Customer Router on the customer site, for any to any connection type

MPLS is a provider tunnel where the customer only sees the ends. What is in between depends on the provider.

On the case of ATM and MPLS, as far I understood it's like "logical" leased line, and how the ISP reserve the bandwidth for the customer ? And what could be the layer 1 protocol ? SDH, SONET ? Other ?

Virtual-circuit switched lines require the provider to reserve all bandwidth in their infrastructure. There is no arbitration for bandwidth, so there's no way to squeeze. Overprovisioning is practically impossible (this is a major reason for the high costs).

The protocol can be anything fitting the SLA - the CPEs provide interfaces for the customer to communicate over, everything else is up to the provider.

Today, practically all lines are Ethernet/IP-based on the customer side. The provider may use any protocol in between and it'd be mostly guesswork to say what's most frequent.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.