The optical carrier CSI

The DS carrier service infrastructure gave us two important building blocks that were used to further extend the  capacity for supporting VoIP networks. First and  foremost, the  DS network established that analog signals could be regenerated in digital  format. Second, the  DS network established that digital  signals could be aggregated with other digitally regenerated signals in the  form of DS0 channels. Thus,  the  capability to channelize digital  bandwidth evolved. Dedicated channels have  proven that they  can support VoIP with
the  same if not  better quality of service that we have  come to expect with
POTS over  the  PSTN.

With the  DS series of standards established, a basis existed for specifying how we might  further scale and  extend bandwidth capacities when  the new fiber-optic cabling carrier services infrastructure evolved. Compared to fiber- optic cables, the  copper-based wiring of the  PSTN and  DS CSIs is much more  expensive to install and  more  prone to failure  due  to electromagnetic interfer- ence, weather, and  the  need to protect the  wiring inside expensive conduits.

Fiber-optic cabling uses laser light and  is not  as vulnerable to these elements. Moreover, fiber-optic cable is more  flexible  and  easier to install. And after  the  use of fiber-optic cable reached critical mass, it became far less  expensive to install compared with nonfiber alternatives.

In 1982, the  first fiber-optic cabling systems were commercialized. That same year, MCI became the  first telecommunications provider company to choose fiber-optic cable to support its national POTS carrier network. Since the 1980s, an entirely new, totally fiber-optic-based infrastructure has evolved. Known today as the  optical carrier (OC) CSI, it followed the template established by the  continuing development of the  DS and  PSTN CSIs. In addition, it further extended the  DS infrastructure by using dedicated and  channelized band-  width techniques. Not surprisingly, the former DS series of standards was used as the  model for determining how to calculate increases in bandwidth thresholds over  fiber-optic cable, how to extend the  geographic coverage areas (including areas not  serviced by the  DS network), and  how to finalize the  standards for OC bandwidth threshold levels  for the  transport services to be provided through the  OC carrier services infrastructure.

When data network standards for LANs, MANs, and  WANs were developed in the  mid-to-late 1980s and  external transports were needed to interconnect various LAN and  MAN sites, both the  DS and  OC carrier services infrastruc- tures were able  to rise  to meet this  challenge. Beginning in the  1990s, carriers elected to install fiber-optic cable whenever possible to supply the  transport demands of their customers. T1 and  T3 lines  formerly based on copper were now being  carved out  of much larger  bandwidth transports of the  optical car- rier  CSI.

VoIP transports go fiber-optic
In the  early  1990s, the  fiber-optic-supported ATM (asynchronous transfer mode) transport service evolved. Before VoIP, ATM was the  only dedicated network type  that integrated data, voice,  and  video applications on the  same network transport. Not long after  the  inception of ATM, some manufacturers developed an ATM option that could be deployed for a LAN solution. But by the  time the  design costs were calculated for the  infrastructure, the  overall cost was higher than any other LAN solution available.

ATM ended up competing with Ethernet, and  Ethernet won. ATM was devel- oped on the  communications side  of the  fence  and  Ethernet was adopted on the  data (computer) network side.  In the  beginning, Ethernet was not  as fast as ATM; it ran  only on slow local area  networks. However, over  time,  Ethernet protocols were adapted to faster transports, such as T1 and  T3. Over these higher-speed transport lines,  Ethernet was more economical because the  equipment to implement it was already in place on the  data network side  of  the  fence.  In addition, the  widespread adoption of Ethernet meant that the necessary equipment became cheaper and  cheaper because of the volume of users. Thus,  the  need for ATM was simply  “passed by” with Ethernet’s faster lines  and  cheaper service.

VoIP runs on Ethernet LANs, and  the  savings from running voice  and video  over  the  same ATM LAN was not  enough to offset  the  startup costs when  compared to Ethernet and  VoIP. Today,  VoIP cost-effectively integrates data, voice,  and  video  on the  same network with Ethernet as the  LAN side  of the  network.

Other VoIP transports
Ethernet is a given on the  LAN side  for any customer implementing VoIP. However, one of the  major decision points for any multilocation company is what  to use as the  transport on the  WAN side  to connect all those locations. In the 1990s, ATM running within the OC CSI had the competitive edge because VoIP was not  yet mature. Today,  this  has  changed. VoIP can run on the  LAN side  and  operate very well with ATM on the  WAN side.  Or VoIP can run on several other OC transport services without the  need for ATM.

Nevertheless, ATM took  off as a MAN and  WAN solution for some companies and  most of the  carriers during the  1990s. Today,  ATM remains the  major transport service used by most network carriers. As a MAN and WAN trans- port service, ATM was hailed as the  superior transport service in terms of quality  of service (QoS), speed, and the convergence of data, voice, and video.  ATM’s quality of service far exceeded the  VoIP alternatives of the  1990s.

At the  same time,  however, ATM costs were high. Early on, it required a mini- mum of an OC-3 transport at each location. (A single  OC-3 transport runs at
155 Mbps  and  is capable of delivering more  than twenty-four hundred DS0 channels.) Because ATM was so expensive, its largest customer base would continue to be the  network carriers, who used ATM to build  their architec- tural presence in both of the  dedicated CSIs (DS and  OS). The carriers used their ATM networks to lease or sell other data, voice,  and video  transport ser- vices  to their customers. Many today use their existing OC service networks to carry  the  emerging traffic from a fast-growing VoIP marketplace.

Any sizable WAN network running ATM service no doubt has  made a large investment in the  cost of ATM-related equipment and  transports. The good news is that  you can run VoIP over such an infrastructure, leveraging the sunk  costs of an OC carrier network against the revenue coming  from carrying VoIP and  other types of traffic. All LANs in your  company have  to be Ethernet, and each LAN needs to be upgraded to support IP telephony.