Of all of the network terms that have vague or even inconsistent definitions, “metro network” ranks high on that...
long list. The problem with nailing down the purpose of metro networks is that shifting missions for the public network have created almost whipsaw changes in the way service providers should optimally connect network equipment.
As more content is stored within a metro area, it makes more sense to get users connected to the content as directly as possible…
Tom Nolle, President, CIMI Corp.
Changing regulatory requirements and competitive forces are also having a major effect. These changes are forcing constant revisions to the optimal infrastructure for the metro network, and that’s likely to continue and even accelerate in the near future.
In the beginning, a local voice call mission for metro networks. In the age of voice calls over the public switched telephone network (PSTN), most calls were local, so switching offices were arranged in a hierarchy to minimize the number of switches needed to complete a given call. This aggregated local exchanges into “metro areas” within which users could make local calls. Each metro area had one or more interexchange points of presence (POPs) that provided inter-metro calling. In this era, a metro in the U.S. was called a “LATA” for “local access and transport area.”
Stage 2: Metro networks morph into broadband Internet aggregators. As broadband Internet grew, new demand brought changes in the way traffic flowed. Nearly all broadband traffic moved through the access and metro infrastructure to an Internet Service Provider’s POP, which meant the role of lower-tier sites or offices in connecting calls was replaced by a role of aggregating traffic for the efficient use of optical transport resources. Because there was little value in providing local connectivity (an email to a neighbor isn’t transmitted between the user pair but through a pair of mail servers that could be thousands of miles away), the access network was often based on tunnels or Ethernet pipes. IP addressing and connectivity was not required within the metro area because routing was then an attribute of the core network,which was separate from the metro.
Metro network mission circles back to local connectivity
The metro network is changing again, thanks to the combination of operator goals to monetize content and handle the increased use of mobile broadband generated by devices, including smartphones and tablets. We’re seeing a renewed need for local connectivity, and that translates into a potentially larger mission for IP devices within the metro network, creating what some call a “metro core” and others call an “expanded IP edge.”
Content, particularly video, is a valuable commodity from a service delivery perspective, but while there are literally millions of elements of user-generated content, the majority of the content that can be monetized is relatively recent TV programming and movies. Because this content is likely to interest a large number of consumers, it makes sense to store or cache it in each metro area where it can be delivered to users with a minimum delay and packet loss risk.
As more content is stored within a metro area, it makes more sense to get users connected to the content as directly as possible, which means an erosion of the notion of a strict aggregation hierarchy and the development of a more connective metro network. While most operators accept that Ethernet aggregation of traffic is cheaper, they prefer IP for more general connectivity.
Handling mobile broadband content. Mobile users and mobile broadband reinforce the trend of placing increased value on metro connectivity. To avoid hauling metro broadband content through mobile trunks and tromboning it for mobile roaming, operators would like to offload Internet traffic quickly. That combines with the need to connect to the best cache given the mobile tower in use and means that the metro network needs to be more IP-based.
Edge of the Internet core moves into metro network territory
The net effect of increased IP in the metro network is to effectively extend the edge of the “Internet core” more into the metro area. Any extension of a core network toward the network edge by moving IP deeper into the metro creates a larger boundary between the core and the access network, and that creates special issues. There are more edge devices, and those devices are likely to be distributed to smaller facilities where power, cooling and space are more limited.
At the same time, at least some of these devices will likely connect to local cache resources using higher-speed trunks, and in many installations the aggregation network will hand off high-speed trunks as well. That means density, power consumption and cooling will be critical in metro implementations. The larger “surface area” of the metro/core edge impacts the Ethernet-to-IP on-ramps, the metro IP routers and even peering points.
The Ethernet-IP on-ramp requirements often include a smaller total number of interfaces per device, and an increased reliance on faster Ethernet trunks with the same features as the older metro edge switches. Because the distribution of traffic in a metro area is not uniform, the largest sources would be broadband-heavy central office locations and 4G wireless backhaul points. Some switches might be quite large, and operators would likely value devices that could be scaled and include smaller and larger platforms to create operational uniformity.
What could be called “metro core” routers would likely be smaller in form factor and capacity than traditional deep-core devices, but they would need the same features, and in some cases, they might even have to scale to the same size as their larger cousins. Existing routers are increasingly used for this mission but were not created specifically for it. The use of metro core routers would be particularly likely in areas with large content data centers or at points where content delivery networks (CDNs) are networked. Operators might also want to peer with mobile broadband roaming partners at the metro level.
There are also device management implications in the trend toward broader IP use in the metro network. A larger IP footprint extends the scope of the IP network in terms of the number-of-devices and number-of-paths, and that could increase the need for effective next-gen management tools and service automation tools for the metro network. That will become particularly true if operators offer any form of enhanced experience to users (read: not best-effort IP services), as many now plan to do.
To sum up the evolution, metro networks of the future, particularly in large cities, could easily surpass entire national core networks of the past in terms of total bandwidth and number of devices. Managing how traffic transitions from the “aggregation” or Ethernet-dominated part of that metro network to the “connective” or IP part will be critical in future-proofing metro network designs.
About the author: Tom Nolle is president of CIMI Corporation, a strategic consulting firm specializing in telecommunications and data communications since 1982. He is the publisher of Netwatcher, a journal addressing advanced telecommunications strategy issues. Check out his SearchTelecom.com networking blog, Uncommon Wisdom.