Niall Robinson is VP of Global Business Development for ADVA Optical Networking.
The expansion of global data centers has been rapid. Driven by the fierce migration to cloud-based services and constantly increasing capacity needs, internet content providers (ICPs) have built bigger and bigger facilities to store and process information.
In the past couple of years though, the trend for ever-larger mega data centers has hit a wall. Once a single implementation grows beyond a certain number of servers, megawatts and square feet, it no longer delivers on new efficiencies and economies of scale. What’s more, as data centers expand, they represent a bigger risk of failure. That’s why cloud-scale operators are focusing on creating multi-data center clusters in particular geographical regions. These groupings of multiple facilities enable ICPs to offer superior business continuity, disaster recovery, and digital media delivery by maximizing the performance, availability, and redundancy of data services.
Of course, high-availability metro-regional clusters can only be effective if the data center interconnectivity (DCI) between them doesn’t become a bottleneck. This creates an urgent need for high-capacity single-span DWDM DCI links. Tbit/s networks optimized for distances up to 80km are required. This DCI infrastructure needs to deliver maximum efficiency in four key areas:
Power: From a sustainability as well as a business perspective, energy usage is as critical in DCI as it is within the data center. Current practice, such as relying on solutions that use active backplanes, often consumes unnecessary watts per bit.
Space: As demand grows, data center real estate comes at an ever higher premium. That’s why it’s critical to leverage an optical transport platform that also minimizes footprint. Data center clusters increasingly depend on ultra-compact solutions with configurations that use the least possible rack space.
Simplicity: With their relatively small operations teams, ICPs require solutions that offer genuine plug-and-play installation and simplified provisioning.
Cost: While equipment flexibility is a high priority for telecommunications operators, when it comes to DCI transport, cost is king. All new innovation needs to guarantee reductions in capital and operating expenditure.
Optimizing and balancing these different considerations is crucial. An important part of this juggling act is deciding between coherent and direct detect methods of optical signal detection. The choice between these two modulation formats has emerged as the key question facing today’s metro DCI infrastructure designers.
In recent years, coherent detection, which involves a local oscillator at the receiver tracking the phase of the optical transmitter, has revolutionized long-haul DWDM networks. It has enabled 100Gbit/s transport on a single wavelength over thousands of kilometers.
Even more recently, coherent detection has been utilized in metro DCI networks. This enhances performance and spectral efficiency, but the increased cost and power consumption make coherent solutions less than ideal for these shorter links. The installation simplicity of coherent system over such short links is highly attractive to the stretched ICP operations teams.
Meanwhile, advances in technology that support 100Gbit/s direct detect DCI applications are challenging the dominance of coherent detection in this market. New solutions built on pulse-amplitude modulation 4 (PAM4) technology are able to transport direct detect signals up to 80km while still meeting stringent optical signal-to-noise ratio requirements.
Here’s how the two modulation formats compare:
- Fiber capacity: Coherent detection offers more in terms of spectral efficiency and, therefore, optical bandwidth. However, the reduced capacity of direct detect is often more than sufficient to connect regional data center clusters.
- Size, power and cost: Direct detect networks are naturally simpler than coherent options, which require application-specific integrated circuit and digital signal processors. This means direct detect systems are cheaper, smaller and consume less power.
- Simplicity: New smart optical line systems, developed exclusively to support 100G direct detect, are leveling the installation simplicity playing field between coherent and direct detect solutions.
- Applications: Direct detect is optimized for shorter reaches and point-to-point connections in metro networks. Coherent technology, on the other hand, is uniquely suited to long distance data transport.
It’s clear why the lower power consumption and lower cost-per-bit of direct detect transmission across distances up to 80km make it a compelling choice for operators of regional data center clusters. What’s more, further direct detect technology research is exploring solutions such as PAM4 with 50Gbaud technology or PAM8. This ensures direct detect will remain competitive to coherent solutions in years to come.
Direct detection certainly has a bright future. It may not be as scalable or as spectrally efficient as the coherent alternative, but its value is increasing rapidly in this expanding space.
Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.