Previous | Table of Contents | Next |
There are users who have network requirements that cannot wait either for stabilized IMA solutions or for high-speed ATM (i.e., DS3 and OC-3c) to become available or more cost-effective in the public network. They need solutions to their networking challenges sooner rather than later. Until the new IMA specification is finalized, they see themselves as having no stable means of interconnecting their ATM networks via inverse multiplexing.
An alternative to IMA allows users to take full advantage of inverse multiplexings NxT1 (e.g., multiple independent T1 connections) bandwidth for their ATM networks. ATM cells can be inverse multiplexed bit-by-bit, meaning that ATM traffic can be transported transparently over traditional fractional T3/E3 and T1/E1circuit facilities. This technique is called clear channel ATM (CCA).
Unlike IMA, clear channel ATM is a specification already approved by the ATM Forum. Clear channel ATM is another way of referring to the ATM Forums transmission convergence sublayer (cell-based TC) specification, also known as ATM over HSSI (high speed serial interface). Cell-based TC specifies a standard format for transmitting cells over any clear-channel bitstream interface. In other words, ATM transports at the bit level, instead of at the cell level. Connectors, clocking, modem control, and status are not addressed by cell-based TC, but are defined by other standards. At the physical layer, V.35, HSSI, or any other type of connector that could accept the ATM bitstream can be used. What is defined by the cell-based TC specification is the bit order and how start-of-cell is determined; then, cells are simply placed bit-by-bit onto the transporting technology (see Exhibit 2).
Exhibit 2. Clear Channel (bit-by-bit) ATM Inverse Multiplexing.
Clear channel ATM means that an ATM bitstream can be carried over any WAN data circuit, including inverse multiplexed data circuits. Having an HSSI port available for the traffic flow is not necessary; a V.35 port will do, and it is even possible to use an ATM DS3 or OC-3c UNI directly from an ATM switch. By using cell-based TC in conjunction with a readily available UNI (e.g., on an inverse multiplexer with a direct UNI connection to an ATM switch), the difficulties and long maturation created by the process of defining yet another UNI are eliminated.
It usually takes time for a new user-to-network interface to migrate into customer premises equipment, appearing most often first in ATM switches, and then in other devices such as routers and network interface cards (NICs). There are no guarantees that a vendor will even choose to support it. In the ATM switch, a new UNI means that the switch must handle a new interface, address the conversion from one rate to another, and buffer cell traffic moving between old and new UNIs. Despite the UNI deployment taking place in the ATM switch, actual switching is not required in this instance. Because ATM switches tend to be expensive, the addition of a new UNI can mean interface upgrades and reconfigurations.
The functions of a cell-based TC imux need not be deployed in a switch because only cell buffering and rate conversion are required. By leaving the switching functions where they belong in the ATM switch a more inexpensive and lower-risk solution can be deployed. In the case of clear channel ATM inverse multiplexing, a readily available ATM interface such as a DS3 or OC-3c becomes an ATM DTE port, which then transports the data stream via multiple T1 or E1 circuits. By converting the traffic from DS3 or OC3 to lower NxT1 rates, and by providing the necessary buffers, the need for a new UNI is eliminated.
Exhibit 3. Private Network/Corporate Intranet: Interconnecting ATM Site Backbones with NxT1/E1 TDM Service.
Previous | Table of Contents | Next |