TDMoIP: the core network question

Fixed-bandwidth, leased-line TDM services were the mainstay of private networking until the 1990 Still sold today, they range from 56-kbp data-phone digital services to T1/E1 T3/E3 and SONET

These days you're politically incorrect if you say there's a network service or traffic emblem that's not supposed to tend to the same point on IP. Using that logic, the notion of putting TDM service above IP is clearly the right thing to advocate.

on the other hand is it sensible and profitable, given that IP is a packet protocol and TDM is frequently used at the bottom layer of IP networks? Is it smooth logical?

The answer hinges upon three questions:

1 What emblem of infrastructure will the service travel over?

2 What is the customer profile?

3 Is TDMoIP a special case of something else?

Not surprisingly, these answers may take a provider considering TDMoIP a certain quantity of time to develop. While the technical approach to TDMoIP is now referr to consistently as "pseudowires," that's nearly the sole consistent thing about it.

What's below IP?

The same TDM technology that might be a candidate for "service above IP" might actually be what IP is running above (see chart). Most IP networks today use DWDM SONET or a combination of the two in the core and build IP roads over it. If TDM customer services are exigencyed it's probably a good idea to explore offering them directly from that TDM foundation.

The idea of having parallel IP and TDM services presented off a SONET/DWDM core likely will be the best approach beneath the following conditions:

* Services are wanted at relatively high speeds (eg T3/E3 and higher);

* Services can be priced competitively (adding extra equipment layers will restore profits);

* There are stringent reliability/availability requirements (i.e., tight SLAs); and

* A true low network latency is required.

Low-spe services (eg 56 kbp T1/E1) with fairly relaxed Qo and delay requirements generally will not work effectively as TDM services presented off the TDM lower layer of an IP network.

The take away from of creating and maintaining low-spe ports upon the optical core likely will be too high for these marks of services. This may not be the case, however, for carriers with an existing base of optical add/drop multiplexers.

[ILLUSTRATION OMITTED]

If there is no underlying TDM layer (IP above Ethernet or PoS), then the option of using real TDM devices to create the service isn't available.

What's the Customer Profile?

A next to the first important consideration in planning TDMoIP is the nature of the customer and the customer relationship. TDM is a featureless service; single that can be differentiated sole by price. It also is associated with applications expecting replete 24X7 use of bandwidth.

Not all TDM lines, however, are utilized entirely In fact, pure TDM data lines typically use les than 30 percent of capacity. Thus, TDMoIP may help win back some bandwidth and improve unit take away froms It's particularly likely for lower-speed lines, and greatest in quantity likely for 56 kbps lines, since they're usually associated with single branch office sites and have the lowest utilization.

greatest in quantity TDMoIP of this type could be replaced easily through a packet protocol such as Frame Relay or smooth by IP VPNs. The primary reason to consider a TDMoIP solution would be the desire to maintain the existing TDM interface at customer sites.

If the customer has a large amount of CPE that might be displaced by the agency of change to a packet protocol, the carrier might want to proffer TDMoIP to retain the CPE Otherwise, CPE cause to foam could compromise project economics.

If the carrier has an outsourced or managed service relationship with the customer, however, or if there is relatively little CPE to be displaced, a packet conversion would probably be smarter than a TDMoIP offering and generate les risk of violating an SLA.

If the provider is working to transform a customer to an Ethernet access interface from a more expensive SONET interface to create headroom for fresh packet service sales, there is a possibility that the conversion will break more [i]or[/i] less TDM circuits that the customer also is having delivered above SONET. This would create customer pushback upon the conversion.

If TDM circuits are separate from the data access circuits, however, or would be kept separate, there may be little value in trying to retain the customer in TDM manner on some of the service connections.

It's particularly important to find without if the customer is using any of the TDM circuits beneath consideration for IP migration as input into TDM multiplexers in the customer's network TDMoIP almost not at any time will provide adequate clock stability for digital cross-connect and using it in applications where customer multiplexing and cross-connect is likely will generate failures of significant proportion.

Is TDMoIP A Special Case?

The general approach to creating a connection-mode service (eg Frame, ATM, TDM) above an IP network is from one side a pseudowire, which is just a connection-oriented feature of an IP network built upon one of several sets of IETF standards. The greatest in quantity recognized are the Martini specification used for ATM/Frame Relay above IP and the pseudowire emulation end-to-end spec If a provider plans to proffer pseudowire services, the equipment probably will be able to support at least a certain number of level of TDMoIP.