Interoute CTO Blames Vendors for Confusion
Delivering a keynote address to vendor and other operators on day one of this year's Ethernet Expo: Europe event in London's Business Design Centre, Finnie noted that the industry debate about the deployment and future of Ethernet should be "a dry, straightforward debate." But it isn't. "For such a mature technology, the level of hype and mis-selling is incredible," he stated.
The debate isn't about the existence of Ethernet as an access service -- it's cheap, everyone understands it, and it's everywhere, he noted.
Finnie is particularly peeved at the complexity introduced by new approaches to carrier-Ethernet technology, especially Provider Backbone Transport (PBT). (See A Guide to PBT/PBB-TE and PBT/PBB-TE Guide: Vendor Talk.)
"We can't understand the tension between MPLS and [carrier] Ethernet -– why is this debate even happening? They are totally different. The idea that Ethernet might replace MPLS isn't happening, in our experience. That debate is being driven by vendors and maybe a carrier," he noted, clearly referring to BT Group plc (NYSE: BT; London: BTA), which is also presenting at the Expo. "The debate is being clouded by specific commercial agendas," he told the audience. (See BT to Update on Ethernet Plans .)
In addition, the introduction of any new technology will introduce extra cost, complexity, and more demands on power and space for carriers that need more simplicity, not more complexity.
On the side of the Expo, Finnie told Light Reading: "The PBT situation needs some clarity. Customers need to be clear about what services they can buy, but the whole Ethernet space is so muddied. Our customers [enterprises and ISPs] get bombarded with bollocks by vendors and other carriers, and most of [the customers] don't really care about the technology -- they just want a service with a service level agreement." But the constant barrage of often conflicting marketing messages leaves many people confused and unsure about what they should be deploying and ordering.
That's not to say that PBT won't be deployed, though -- Finnie noted that it's a case of "horses for courses," and that decisions will be made depending on the history of network operators. Some, he notes, might like the idea that PBT is supposed to have operational aspects that are similar to SDH, for example.
And he can see PBT being used at the edge of networks to help manage retail, consumer services. "I can see it being deployed as a service control point mechanism, where a carrier needs to hive off bandwidth for a specific application. It makes sense in that context," says Finnie.
Packet transport costs falling
Interoute sees networks evolving to a simplified, converged architecture, based on an OTN [Optical Transport Network] core, managed by GMPLS, and MPLS, with Ethernet as one of many available services. And because Interoute already has fiber in the ground around Europe that it can light, it is "using capacity to avoid complexity."
And that approach is becoming a more cost-efficient way to handle service and bandwidth demands, he noted. "Demand is currently being driven by consumer broadband. [As a result] there's an insatiable demand for IP -- we are adding an additional 780 Gbit/s of capacity this year -- and greater demand for network capacity. By the end of 2008 we will have grown our network capacity to 5 terabits per second, more than double at the start of the year." Broadband growth in Eastern Europe is a major driver of new bandwidth demand for Interoute, notes Finnie.
That's affordable now because of the "optical step change" that occurred in 2006 with the introduction of photonic integrated circuit (PIC) technology -- Interoute is a major customer of PIC-based optical equipment vendor Infinera Corp. (Nasdaq: INFN). "Technical capabilities have changed the economies of optical transport," stated Finnie. (See Infinera Muscles Into Interoute.)
And now that same trend is affecting the packet-transport world. "We are seeing a compression of price in the router world too. The difference in price between optical and packet transport is shrinking." And it's about time too, reckons Finnie, who believes the cost of packet transport systems has been artificially high for too long.
"In the past there's been no real reason why a service based on Cisco Systems Inc. (Nasdaq: CSCO) or Juniper Networks Inc. (NYSE: JNPR) [technology] has been more expensive than a service based on Ciena Corp. (NYSE: CIEN) or Infinera. It was more expensive because it could be," said Finnie.
— Ray Le Maistre, International News Editor, Light Reading