Shortening the (DAA) Last Mile

"Where other network architectures like wireless networks have gradually pushed fiber closer to the cell tower and increased bandwidth as demand has grown, DAA requires an unprecedented step change in access architecture and bandwidth that brings significant challenges." —Jon Baldry, Infinera

Jon Baldry, Infinera

Jon Baldry, Infinera

The cable industry is buzzing with the prospect of distributed access architectures (DAA) and the resulting step change in end-user experience. DAA is a once-in-a-generation architectural change that will enable cable MSOs to defend their competitive advantage in terms of broad content offerings built upon industry-leading bandwidth per user.

DAA has many facets, but from a data aggregation and backhaul perspective, the biggest change is the fiber-deep push within the access network. Migration to Wavelength Division Multiplexing (WDM)-equipped Remote PHY devices (RPDs) requires the previous WDM-based transmission network to extend from secondary hubs deeper into the access network, connecting directly to the RPD.

This sounds like a simple proposition — WDM has pushed deeper into access networks over the last decade as fiber availability has increased and technology costs have dropped. Why shouldn’t DAA be another example of this, with a gradual push for WDM deeper into the access network to support the rollout of RPDs?

The reason it is not is the significant increase in scale that DAA drives into the underlying transport network. Where other network architectures like wireless networks have gradually pushed fiber closer to the cell tower and increased bandwidth as demand has grown, DAA requires an unprecedented step change in access architecture and bandwidth that brings significant challenges.

If we look at a typical residential area served by a cable MSO, the secondary hub serves clusters of approximately 500 homes passed with an optical node or optical network termination and a corresponding amplified hybrid fiber coaxial (HFC)-based access network with five or more HFC amplifiers per network. To achieve DAA, 10 or more RPDs could be deployed within this network to remove most if not all these HFC amplifiers. Looking at the U.S. market as a case study, we can understand the number of RPDs that will be required. U.S. cable networks pass around 122 million homes, and as RPDs typically support 50-60 homes passed, a simple calculation gives us a market of around 2 million RPDs, each with a new 10G DWDM backhaul link into the transport network. Unfortunately, residential areas haven’t always been built in convenient clusters of 50-60 homes, so the actual number of RPDs required in the U.S. may be closer to 3-3.5 million.

Those figures address the macro-level requirement for new 10G DWDM access circuits driven by DAA. At a more local level, secondary hubs vary in size, but the largest will need to support 600-700 RPDs, and more in cases of secondary hub consolidation. We therefore have a twofold challenge--a massive rollout of new 10G DWDM access circuits across cable networks, and huge concentrations of these circuits within secondary hubs that are often already space and power constrained.

Looking first at the hundreds of thousands of 10G circuits that will hit the DAA access network, we must consider how to address the operational aspects of the rollout. The key requirement is simplifying the rollout process through automation. Autotuneable WDM-PON technology that allows the network to tell the remote optic which wavelength to tune to automatically has been around for several years, but with limited bandwidth and reach. Recent technological advances have extended both capacity and reach, making this an option for DAA and other fiber-deep deployments. Autotuneable DWDM optics deployed within the RPD will enable deployment engineers to treat them like grey optics, without the need for knowledge of WDM or the exact WDM wavelength requirements of the specific location. This will greatly simplify the process, reduce spares holding costs and decrease the likelihood of deployment errors.

Returning to secondary hubs, again operational issues need to be foremost in our minds. Key issues to consider in the secondary hub are:

  • Space and power – real-world deployment scenarios are often space and power constrained, and DAA transport and aggregation solutions must capitalize on available space and minimize power consumption. 
  • Fiber management – Many secondary hubs will terminate 600+ RPDs, meaning 1,200+ fibers connecting the WDM filters to the aggregation platforms before other connections within the rack are considered. Scalable DAA transport solutions must address this critical issue.
  • Automation – DAA will capitalize on the most advanced automation and network optimization trends, such as autotuneable optics to the RPD and the central office rearchitected as a data center program. DAA aggregation switches must support these initiatives to ensure that MSOs can keep ongoing operational costs down while allowing networks to rapidly scale.

DAA will allow MSOs to fight off alternative access providers and keep their competitive edge. It brings considerable challenges throughout the network, and those within the transport and aggregation domains shouldn’t be underestimated. The good news is that the optical networking industry is innovating to address these operational challenges, allowing the transport network to be rolled out quickly and simply, then rapidly scaled to deliver on the promises of DAA.

JonBaldry is director of the metro business unit of Infinera, an optical networking solutions company based in Sunnyvale, Calif.