Optical/IP Networks

Riverstone Aims for Access

Augmenting its service provider pitch, Riverstone Networks Inc. (OTC: RSTN.PK) today announced products that extend the company's carrier Ethernet reach to the access network.

The new 15100 and 15200 series are smaller versions of the flagship 15008 router (see Riverstone Adds Ethernet Routers). Yes, vendors are always releasing smaller versions of routers, but Riverstone says these new boxes are meant to fill an industrywide gap in carrier-class Ethernet.

Carriers that put Ethernet in their networks have been using enterprise-class routers in the access network, says Inbar Lasser-Raab, Riverstone vice president of marketing. Hence, the company says its 15100 and 15200 lines bring carrier-class Ethernet down to a box suitable for access networks or even customer premises. "With these platforms we allow carriers to move OAM [operations, administration, and management] capabilities into the access," Lasser-Raab says.

Some routers follow a similar philosophy -- the 3750 from (Nasdaq: CSCO) can put Ethernet and Multiprotocol Label Switching (MPLS) at the customer premises, for example. But the 3750 was derived from Cisco's enterprise routers. Riverstone claims it stands out by having the first service-provider-oriented boxes targeting this space.

The new boxes use the same software as the 15008 and provide the same features. The difference is the size -- the new versions are two rack units high, with a maximum of 20 ports of Gigabit Ethernet. The 15100 line uses four ports of Gigabit Ethernet as uplink, while the 15200 boxes provide two ports of 10-Gbit/s uplink.

Riverstone has announced two customers for the new routers: Updata Infrastructure UK Ltd. and the Utah Telecommunication Open Infrastructure Agency (UTOPIA). Carrier Ethernet is gaining favor quickly in some circles, aided by the Metro Ethernet Forum (MEF)'s formalization of requirements to be "carrier-class" (see MEF Rubber Stamps Ethernet Gear and MEF: Certification Wasn't Easy). That the idea is percolating further to the edges of the network is no surprise. "Carrier Ethernet is beginning to play in every part of the metro and access network infrastructure," writes Heavy Reading analyst Stan Hubbard in a recent report, "Carrier Ethernet Equipment Market Outlook."

— Craig Matsumoto, Senior Editor, Light Reading

Pete Baldwin 12/5/2012 | 3:01:08 AM
re: Riverstone Aims for Access Is this a good direction for Riverstone? What would you do to bolster the company's future?
Polo88 12/5/2012 | 3:01:05 AM
re: Riverstone Aims for Access In order to improve the company's future, Riverstone must file its remaining outstanding financial restatements, resolve the formal SEC investigation, and get its stock relisted back on the NASDAQ!! As far as their new products are concerned, its not a bad idea to extend Ethernet into the access networks.

Ethernet is a mature, well-understood technology. Due to its wide deployment and broad market support, Ethernet is a low-cost technology, both in terms of equipment costs and operations and management costs. Unlike SONET/ATM, Ethernet has a volume of commercial deployments driving its cost and technology evolution. Likewise, Ethernet and IP are easy to provision, and a large pool of technicians exists.

High bandwidth in scalable increments is another compelling advantage of Ethernet. During its more than 25-year life span, Ethernet's bandwidth has increased from 10 Mbps up to 10 Gbps, with 1 Gbps interfaces readily available. In addition, the Ethernet 802.3ad link aggregation standard allows operators to increase the bandwidth between connected devices by logically combining multiple links into a trunk.

The two main different technical approaches taken by FTTP vendors today are PON, and Active Ethernet. PON is current much more widely deployed by some of the world's largest incumbent carriers (including all three of the RBOC's) compared to active Ethernet since it has already been standardized, and is perceived to be easier and cheaper in terms of integration into their current legacy (ATM, FR, and TDM) and next generation IP network infrastructure.

PON supporters argue that competing technologies, such as Active Ethernet, simply don't fit the access world. A point-to-point architecture isn't the most cost-effective, interoperable, scalable and manageable way to get multiple services to a large number of geographically diverse users.

Standards-based GPON, on the other hand, fits perfectly with the business model most carriers have planned for providing multiple services to multiple customers. With PON, the service provider lowers the cost for high-speed, high-bandwidth services by using a shared optical infrastructure. Each service - voice,video and high-speed data - has unique performance characteristics that require different bandwidth allocations and capabilities. While this could pose significant challenges to competing technologies, a GPON system is designed to meet the diverse QoS needs of current and future service offerings.

The following is courtesy of:


Proponents of Active Ethernet such as Riverstone argue that Active Ethernet requires an individual optical fiber connection per user. Each user is independent of the other and any combination of distances maybe served.

This user independence also serves another purpose. Today business parks with multiple tenants may require more than a 100Mbps of bandwidth, and using the same fiber a Gigabit Ethernet or 10 Gigabit Ethernet connection maybe provided by provisioning the appropriate port on the Ethernet router. In the future, other locations which require Gigabit Ethernet speeds can undergo the same transformation easily.

PON on the other hand, makes each user dependent on the other users, since it is a shared medium. So the optical budget of the PON has to be divided among the users of the PON. Therefore the service provider has to have early and accurate knowledge of the take rate in its serving area on a granular basis and the distance of each of its users. While distance to each of the users it plans to serve can be determined, take rates on a granular basis are historically hard to predict. A mistake in the take rate and user density assumptions could result in severe penalties with a PON.

This is because PON is a shared medium and typically uses a two-stage splitter architecture where each OLT is connected to a 1:4 splitter, followed by 4 1:8 splitters to accommodate 32 users. This limits the reach of the fiber from each of the splitters. So even a small serving area has to be wired for a expected user rate and any variation results in increased costs or lost revenue. Under-predicting the take rate results in users who cannot be serviced without more investment; over-predicting the take rate results in investment that is unused. Neither of these options results in success for a carrier in a competitive environment.

An Active Ethernet architecture typically uses an aggregation Ethernet router located in a street cabinet closer to the end user. PON uses a passive splitter at this aggregation point. The use of a passive splitter is listed as a capital cost advantage for PON since it does not require electrical power and has a small form factor. However, this difference in the number of active elements is limited to the aggregation routers in the field.

PON requires power at every customer premise just like Active Ethernet does.
PON requires power at the OLT just like Active Ethernet does.

With Active Ethernet, in large FTTH deployments, up to a 1000 users can be aggregated into one street cabinet. Thus even for a deployment with > 100000 homes, the number of additional locations where power is required with Active Ethernet is in the 100s. This high aggregation ratio can be achieved because Active Ethernet supports greater distances than PON.

Active Ethernet also offers other benefits in the longer term which decrease operating expense; increase revenue potential and quickly payback the additional investment in powered field locations.

Sign In