Discussions

Based on the analyzed results, the MAN access strategy “A” focusing on the SONET/SDH and ATM technology has the highest BNP score. This result highlights the following points:

1. Although SONET/SDH system has limitations, its ability to provide quality of service and network surveillance information outweighs its technical shortcomings, such as bandwidth scalability.

2. Since the evaluator from the telecom carrier group is an incumbent carrier in Taiwan, this result reflects that the incumbent carrier prefers a path of evolution rather than revolution.

In other words, leveraging the existing SONET/SDH infrastructure to offer broadband services is more important than eliminating the existing SONET/SDH systems and deploying a new Gigabit Ethernet network. The issue of total network cost may prohibit incumbent carriers to have a “desert start” in deploying the Gigabit Ethernet network in the MAN environment. On the contrary, for a new telecom carrier, because they do not have the existing network infrastructure compatibility issue (i.e., a “Greenfield” deployment), the choice of network architecture and access technology typically will favor the promising technology with good potential which, in this case, would be the Gigabit Ethernet technology. This is consistent with the approach that most of the telecom carriers, such as Yipes (www.yipes.com) and OnFiber (www.onfiber.com), serving the metropolitan customers have selected Gigabit Ethernet as their access technology.

3. The weight of network operations aspect (0.337) is much higher than the weight of network equipment cost aspect (0.241). This result indicates that telecom carriers should focus on the total cost (first cost and future maintenance cost) of the network equipment and not just the initial equipment purchasing cost. Particularly, if the incumbent telecom carrier already has the network management systems to support the circuit-based SONET/SDH operations, deploying the IP-based Gigabit Ethernet technology will create challenging network

operation compatibility issues for the telecom carriers. If the telecom carrier decides to operate 2 layers of network operations, then the total network operation cost may offset the Gigabit Ethernet equipment cost advantage. How to cost effectively manage the access technology will be particularly critical to the incumbent carriers when offering broadband service.

4. The purpose of technology is to provide service to users. The first priority of a telecom carrier is to offer high quality service to its customers. Therefore, a telecom carrier should be “service focused” and not only focus on technology.

5. Bandwidth scalability criteria (0.169) received the second highest score of all criteria. Since bandwidth scalability is one of the major strengths of Gigabit Ethernet technology, if the issues of QoS, network management functions, and network compatibility can be addressed, it is conceivable that the Gigabit Ethernet technology will play a major role in the future broadband network.

6. Strategy A and strategy B BNP values were 92.49 and 80.24, respectively. These 2 BNP values are relatively high scores compared with the possible 100 total score. They reflect that both access technologies have the capabilities to address different carrier’s network requirements, particularly strategy A. In addition, the BNP difference between the two strategies is only 12.25, possibly suggesting that a combination of the two strategies could form an evolutionary approach as a viable way to address the MAN access technology issue.

Since June 2002, 10-Gigabit Ethernet technology has been standardized to be compatible with the OC-192 SONET/SDH transport payload format. Therefore, the deployment of 10-Gigabit Ethernet technology over OC-192 SONET/SDH (Ethernet over SONET/SDH) infrastructure appears to be a promising alternative for future broadband service (IEEE Standard 802.3AE-2002). It is observed in Table 3 that working with the existing

infrastructure (backward compatibility) while evaluating a new network technology is an important issue. Therefore, the carrier’s position - incumbent or start up - will have a big impact on the selection of a new network technology.

From the results of practical applications in evaluating the two access strategies, the proposed method makes a good evaluation and appears to be the most effective and appropriate one in a fuzzy environment. Through this fuzzy MCDM method, senior business executives could gain significant insight on the telecom carrier’s network and business qualitative issues while setting a MAN access strategy.

4.4 Summary

Ethernet technology has evolved into multi-gigabit bandwidth with applications extending from LAN, MAN, and even to WAN. Telecom carriers are faced with fast growing Internet data traffic that exceeds traditional voice traffic. Gigabit Ethernet technology holds the promise of relieving the critical bandwidth bottleneck in the access network. Although the simple Ethernet protocol can offer telecom carriers many advantages in terms of simplified network architecture, substantial equipment cost reduction, and dynamic bandwidth allocation flexibility compared with existing SONET/SDH and ATM infrastructure, Ethernet technology also possess weaknesses in the areas of network operations and management. How to select the appropriate access technology in the MAN becomes a critical network planning issue, particularly in light of emerging broadband service demands.

Because a lot of technology insight is not quantifiable, how to “measure” these

“qualitative” issues, such as a technology’s compatibility with existing infrastructure is of particular importance to telecom carriers. Depending on a carrier’s business position (such as incumbent carrier versus new start-up), different approaches may be applied by different carriers when setting an access technology strategy. This study discusses how to apply the fuzzy MCDM method to evaluate different access strategies in a fuzzy and complicated

network and business environment. Using this method, not only can telecom carriers avoid overlooking significant qualitative issues, but they can also prioritize network and business objectives while establishing their access technology strategy.

Lastly, our example of an empirical study in Taiwan is based on the results of a generalized model evaluating the network access strategies in a fuzzy environment. From the results of this practical application, it was found that technology advantages alone may not be sufficient to address all business and network issues. This proposed method appears to be appropriate and effective to address the qualitative issues in a fuzzy business environment.