Feedback

Beyond the city limits: How Ethernet is exceeding the Metro boundary

EnterpriseFeatures
by James Morris, 11 Jun 2014Features
Beyond the city limits: How Ethernet is exceeding the Metro boundary

Ethernet used to just be the technology for the local area network. Once you reached the boundaries of a particular building, a different technology would need to be employed for connections between distant buildings and to the outside world in general.

For small businesses, this was ISDN, then ADSL, SDSL, VDSL and cable modems. But larger businesses have traditionally relied on leased lines, in particular T-carrier or E-carrier technology, depending on whether you are in the US or Europe. These bundle Internet connectivity with telephone services, using the infrastructure of public telephony to provide the necessary data connections.

However, the leased line is very old technology. T-carrier dates back to 1957, and although T3 lines can handle up to 45Mbits per second, and E3 lines 34.4Mbits per second, they are no longer the connection of choice. With the advent of fibre, technologies like Synchronous Optical Networking (SONET) in the US and Synchronous Digital Hierarchy (SDH) in the rest of the world have provided higher-bandwidth trunk connections.

But SONET/SDH can be prohibitively expensive, due to its origins as a solution for large-scale circuit-based communications, and the need to translate the packets of Ethernet local networks into SONET/SDH frames. This has pushed fibre-based connectivity well into the realm of service providers and large enterprises only.

Nevertheless, there is a growing hunger for high-bandwidth connectivity. The cloud is a significant driver. The need to access files and even software on demand from a remote location is constantly on the rise. Software as a service (SaaS) is increasingly important, as well as online audiovisual material being used for training, voice over IP, and webinars replacing physical meetings.

Growing interest in ethernet

All of these activities require high bandwidth and low latency. For businesses that make extensive use of cloud-based services, reliably fast connectivity is therefore essential. Heavy use of remote desktop VPNs will hammer a network, too, and if your business relies on a shared database, instantaneous synchronisation when records are updated will keep employees productive.

As a result, there has been growing interest in Ethernet as an alternative technology for operating wide area networks (WANs) over metropolitan areas, known as Metro Ethernet. Although 1, 10, 40 and 100 Gigabit Ethernet supports distances of up to 100m over copper wiring, it can be delivered up to 40km over fibre. This is more than enough to cater for even quite large metropolitan areas.

A Metro Ethernet solution fulfils the need for bandwidth extremely well. It is already about a third of the price of providing the same bandwidth via leased line technology. But it is also capable of providing far more bandwidth than leased lines, and potentially more than SONET/SDH as well.

It can be supplied in an instantly scalable form, so if there is a spike in bandwidth requirement, this can be catered for temporarily. Metro Ethernet fulfils the needs of connecting resources in a metropolitan area with higher bandwidth and lower costs than previous technologies.

But the potential goes way beyond just wiring up key locations in a city. Metro Ethernet is becoming a high-bandwidth option for all manner of connectivity, upgrading a host of legacy technologies and providing the backbone for future applications too. Its latest developments give it more seamless connection to the Internet, the potential to play a key backhaul role for the most recent wireless data standards, and a more flexible selection of service topologies. So it has the features and opportunity to go well beyond the leased line replacement role it took in its original incarnation, empowering the next generation of networked activities.

Carrier ethernet

The key factor with Metro Ethernet is that it is essentially Ethernet. It employs a special version called Carrier Ethernet (CE), but the protocol is exactly the same, so it benefits from the lowered costs afforded by being able to use ubiquitous commodity hardware and expertise for large parts of its implementation. It is completely compatible and so doesn't require any special equipment or configuration, as no signal conversion is required. This is a key differentiator from SONET/SDH. Allied with the uniform protocol is greater security, as there are fewer opportunities to intercept communications as data packets are converted between formats.

Carrier Ethernet has been adapted for effectiveness over long distances, however. Fibre is the main conduit, and the CE 1.0 specification standardised the way Ethernet is implemented over this conduit, bundling together standards proposed by the Metro Ethernet Forum (MEF). Providers of Metro Ethernet will often use the same fibre network cabling that they use to provide mobile phone tower backhaul. This is what makes upgrading capacity for corporate clients so easy, as the capacity is already available in the infrastructure. Just a few adjustments to client equipment are required to increase bandwidth.

A significant enhancement came with the arrival of the CE 2.0 in 2013, however. The first version of the CE specification standardised Metro Ethernet for delivery over one provider's network, detailing two main services – E-Line and E-LAN – and a variety of MEF technical specifications for setup and management. CE 2.0 adds E-Tree and E-Access services, alongside a host more MEF technical specifications for additional attributes and more elaborate management scenarios. And whilst CE 1.0 was focused primarily on standardisation, CE 2.0 allows multiple classes of service (Multi-CoS), with managed and interconnected characteristics.

One of the key features amongst all these standards is more efficient bandwidth utilisation, where different classes of service can be given a hierarchy of priorities, which are maintained across the various stages of the connection. So, for example, real-time voice communications can be given priority over video, and both given priority over data. These can also be bundled and transferred across providers, the relevance of which becomes clear when you look at the new types of service that have been added with CE 2.0.

Of the original CE 1.0 services, E-Line is intended for point-to-point connections, with support for port-based Ethernet Private Lines and VLAN-based Ethernet Virtual Private Lines mulitiplexed services. So it is aimed at connecting two geographically distant office networks, or a business to a service provider. E-LAN, in contrast, is multipoint-to-multipoint and offers Ethernet Private LAN and Ethernet Virtual Private LAN variants. Since it can multiplex multiple Ethernet Virtual Connections at each user-to-network interface, it can connect a number of locations and even transport different, discrete networks across the same infrastructure.

How E-Tree service differs

The New E-Tree service type, however, is a rooted point-to-multipoint system with roots and leaves. Whilst roots can exchange data with each other and with leaves, the leaves can't exchange data with each-other. This is aimed at providing traffic segregation for cloud services, media distribution and franchise networks. For example, you could have multiple load-balanced cloud servers that can work together as roots, which the users can access these as leaves without being able to access each-other. As with the other types of service, there are regular Private and Virtual Private flavours of E-Tree.

Finally, and perhaps most significantly, E-Access is about connecting different provider networks into one seamless whole. A service provider partners with an access provider to reach a remote customer location. The data is tunnelled through the access provider's network to the customer. This service is based on the MEF 33 standard, which simplifies the interconnection of the disparate services.

E-Access in particular relies on the CE 2.0 Multi-CoS ability, because it preserves classes of service for the subscriber, including bandwidth profiles and the Service Level Agreement, whilst transferring the priorities across the single E-Access class of service. As there are Private and Virtual Private flavours of E-Access, this makes eight different service types defined in total within CE 2.0 (two flavours of four basic types).

The CE 2.0 specification also offers better integration with the wider Internet. Hosted apps can be integrated with the Internet over a single connection, for example making public-private cloud hybrids more seamlessly usable. There's support for 4G/LTE migration, so Carrier Ethernet can be used as backhaul to connect the many cell towers to the backbone network with enough bandwidth to support the huge leap in data demand of 4G/LTE compared to 3G.

Here again the Multi-CoS abilities will mean that particularly latency-sensitive data types like voice and streaming video get the service they require for end-user satisfaction.

The CE 2.0 specification radically opens out the range of options for Metro Ethernet, pushing it well beyond the original concept. Where previously it was great for providing high-bandwidth private lines between metropolitan locations, either point-to-point or multipoint, now it can be used for a whole lot more.

Broadcasting to end users over Ethernet is now facilitated by the E-Tree service type, and Metro Ethernet can be provisioned wholesale by service providers thanks to E-Access, so companies won't have to rely on one vendor to provide infrastructure in every location. This will accelerate adoption beyond merely providing fast connections between locations within a city, to regional and even national applications.

This post is brought to you by Comcast Business. Follow us @comcastbusiness.

Topics
blog comments powered by Disqus