New-look SDH to the rescue: next-generation SDH can help operators cut costs and increase revenue in the metro area network - The Bandwidth Economy - SDH protocol - Industry Overview

Telecommunications International,  Nov, 2002  by Steve Saxon

• E-mail
• Print
• Link

Data and IP traffic in metro networks is not only increasing exponentially but it's becoming more dynamic -- a mix of different protocols and the need for ever-changing data rates has seen to that.

So what is the best strategy to handle growing demands for higher bandwidth, varied protocols and new services? To succeed -- and maybe even to survive -- carriers need maximum efficiency and flexibility in their operations in order to meet traffic volumes and customer demand. They need to slash bandwidth-provisioning time as well as to shorten the period needed to install new revenue-generating applications. And, of course, they're under huge pressure to reduce capital and operational expenditures while maximising current network investments to the full.

Cost-conscious network operators face a dilemma. They have invested millions in SDH equipment but to meet current and future demands they need to move beyond the capabilities of existing networks -- but they don't want to replace their existing SDH network.

The good news is that they don't have to. By 'enhancing' their existing network with new equipment in the access and the metro, operators can find the most economic route to transforming their traditional SDH network into a platform capable of meeting today's and tomorrow's demand for voice and data traffic, varied protocols and flexible bandwidth. And since it means extending the current network rather than a huge overhaul, the cost is relatively modest.

SDH today

SDH is the dominant transport protocol in long-haul networks (voice and data) as well as in voice-dominated metro networks. As a resilient, well-understood transport mechanism, SDH has stood the test of time. It offers carrier-class reliability -- 99.999 per cent availability based on 50ms switching to backup paths, combined with extensive performance monitoring. SDH also provides transport for different traffic types, whether voice connection, ATM or P with each protocol supported by the latest QoS capabilities.

But current SDH networks have shortcomings that create challenges for network operators. Traditional SDH was designed to carry voice traffic and, as such, is a rigid and inefficient method for transporting data. It offers relatively coarse granularity derived from the old voice networks -- 2Mbps, 34Mbps, 165Mbps and so on. Full bandwidth is required to handle bursts of data traffic even if connections are usually empty or not used to full capacity. The unused bandwidth provides no benefit yet the customer still has to pay for it.

To compound these inefficiencies, service provisioning is a time-consuming and costly business. As it's a manual process, adding new bandwidth can take days and may require an onsite visit by a technician. At a time when carriers more than ever need flexibility and speed in adjusting bandwidth to meet customer demand and so gain a competitive edge, this cumbersome provisioning process is a burden that is increasingly difficult to justify.

Is optical ethernet an alternative?

Metro networks for data services can be based on optical ethernet equipment, an evolution from ethernet's traditional LAN role to cover longer distances and more extensive networks. In ease of configuration, ethernet holds the advantage over traditional SDH, including fine granularity to closely fit a customer's bandwidth needs. Also, unlike SDH, ethernet architecture is designed to handle bursty data traffic, providing a more efficient approach to allocating bandwidth. As a result, ethernet is moving beyond building and campus settings to support organisations -- such as universities or hospitals -- that may have facilities throughout a metropolitan area. In this scenario, the organisation will maintain two separate networks: SDH for voice and ethernet for data.

However, ethernet has drawbacks of its own. In contrast to SDH, ethernet was designed for data not voice (or other streaming traffic such as video). A fact of life with ethernet transport is packet collisions, requiring that packets be re-sent whenever a collision occurs. While delays caused by packet collisions may not matter when sending static data such as text e-mail, the consequences can be severe for real-time traffic such as voice or video. Ethernet vendors are currently working with the IEEE on drafts to create a new standard to support this streaming traffic. However, the results are still unproven and no equipment has been released to market.

Moreover, ethernet has not achieved the reliability necessary for carrier class transport -- uptime in the 99.999 per cent range. Manageability, especially performance monitoring, remains a problem. For example, if a link is broken the system will report that overall network performance is degraded. Unfortunately, it won't provide the other crucial bits of information required -- the fact that the problem was caused by a broken link in the first place and where it occurred within the network. Consequently, ethernet networks become unmanageable once they exceed a certain number of links.