Turning old into new
By Thomas G. Robinson, Executive Vice
In the networking world, though, there are often opportunities to turn the old architecture into something new and useful. For those who prefer complete network overhauls or forklift upgrades, such "old into new" pursuits might be seen as employing "slipcover" technology. This, however, would imply that the original network has lost its functionality. A better perspective is that the original network continues to have a solid foundation, but cannot achieve the full functionality required to meet current and future needs without modifications. An excellent example of this concept is the large scale deployment of HFC networks, where the "C" largely remains from the existing network and the "F" and upgraded electronics are added to enable the new/old system to meet current and future needs. It certainly is far more than a mere slipcover, but avoids the high cost of a complete overhaul such as development of a fiber-to-the-home (FTTH) network.
On the business and institutional side, there are a number of options to make the old new. For example, some organizations that have fiber-based Wide Area Networks (WANs) in a traditional star topology have found a need to now develop redundancy in the network. However, they can't at this time afford to install two discrete paths to and from every site or rebuild their topology as a large, redundant ring. One of the options in this case is to develop rim trunks between neighboring sites. This will enable at least an alternate route and can facilitate the development of multiple, interlocking, smaller rings by adding a relatively small amount of new fiber infrastructure.
Once implemented, the rim trunks will enable redundant paths from any of the sites. These paths, depending on budgetary considerations and how traffic is combined on the alternate path, can be established through: an always available, bi-directional ring; an automatically switched backup circuit using hot standby electronics; or a redundant path that is manually switched when needed.
In other cases, a limitation on the WAN may be the available fiber capacity remaining from the original WAN infrastructure, especially if a number of pairs were initially dedicated for specific, rather than integrated, applications. For example, for each site, one pair may have been dedicated for data; another for T-1 voice connections; and another for analog video. Now, however, the IT staff has decided that additional services are required, but additional fiber capacity to implement them in the traditional manner is not available.
Two main options exist here. One is to combine all services by digitizing and potentially also packetizing them such that they can then be transported using an integrated transport technology such as ATM, gigabit Ethernet or another proprietary technology. If a ring can be cost-effectively configured, potentially Sonet and transport methodologies on the horizon, such as Resilient Packet Ring (RPR), could also be implemented. The other option is to continue utilizing the traditional dedicated carriage methods, but employ wavelength division multiplexing in either a simple or dense fashion.
This will enable the existing services to be maintained in their current form, and new services to be added in a cost- effective manner without having to change the underlying type of transport or constructing additional fiber. Because the recent trend for both of these methodologies has been to come down in equipment cost, 2002 may end up being a good year for network managers to maximize their existing fiber infrastructure.
Other candidates for network modification rather than complete overhaul are some of the old all-coaxial institutional networks (I-Nets) that may have sat dormant for years or seen only minimal use. For a number of these I-Nets, current applications may have leapfrogged the data, voice and video carriage capabilities of RF systems such that optical networking is now required. However, with the strides that have been made with RF broadband-based communications, including the latest version of DOCSIS, the advent of PacketCable and recent innovations with broadband Ethernet systems that would enable 100 Mbps or greater transfer over an RF-based system, there are present and planned I-Nets that can benefit from an upgrade to an HFC infrastructure. Experience indicates that dormant I-Nets would need to be completely rung out, reconfigured and upgraded such that the fiber portion of the system runs deep into the network to dramatically reduce amplifier cascades. In some cases, a hybrid optical/HFC I-Net would also need to be developed. In this case, the HFC nodes could have excess fiber infrastructure and be located at or near high communications capacity institutions. This would enable these institutions to have pure fiber connectivity, while institutions further down the line could be connected to the lower capacity HFC network.
Whatever your networking scenario, my desire for all of us is that we encounter a safer and happier new year in 2002.
|Have a comment? Contact Tom by e-mail at: email@example.com|