x
Optical/IP

Marconi Scales Down Its Edge

Marconi has joined the multiservice edge aggregation party with the launch of a customer premises box that offers a mix of SDH and Ethernet ports (see Marconi Launches New Edge Box).

It's the latest in a string of products that combine multiple technologies in a single, unobtrusive unit. And instead of joining Lucent Technologies Inc. (NYSE: LU) in announcing its product in time for last week's Supercomm show (see Lucent Joins the Edge Crowd), Marconi held back its launch for this week's more European-centric WDM and Metro Optical Networking Conference in Cannes, France. Which makes sense: Marconi's customer base is strongest in the SDH markets of Europe and Asia/Pacific, rather than the Sonet world of North America.

So what has the British vendor come up with? Basically, it's a compact, pizza-box unit (4.5cm x 45cm x 21cm) with two STM1/STM4 ports for SDH, four E1 ports for 2-Mbit/s Ethernet connections, and four Fast Ethernet (10/100-Mbit/s) interfaces. Matthew Smith, the vendor's SDH product line manager, says the box incorporates the next-generation SDH features carriers are demanding for transporting Ethernet traffic, specifically GFP (Generic Framing Procedure), LCAS (Link Capacity Adjustment Scheme), and virtual concatenation.

The idea is that Marconi sells the box to operators, which then install the product at their enterprise customers' buildings as they start using Ethernet services, which are growing in popularity (see Europeans Go Crazy for Ethernet Services). Smith says Marconi already has some orders for the OMS840.

"We already have edge equipment that includes Ethernet, but this delivers a mix of Ethernet and TDM interfaces in a smaller, cheaper footprint," says Smith (see Gearmakers Chase Ethernet in Europe and MRV, Marconi Team on EoSDH). "It's aimed at PTTs and alternative operators, which are all seeing high demand for Ethernet services now from their enterprise customers. The carriers wanted a less-expensive device to sell into their business users," he adds.

That's exactly the message Cisco Systems Inc. (Nasdaq: CSCO) is hearing, too (see Cisco Renews Its SDH Chase). Marconi, though, is not prepared to provide any pricing information for the OMS840, so it's unclear just how cheap this product is.

Heavy Reading Chief Technologist Geoff Bennett says it's about time Marconi launched such a product, especially as it was one of the first vendors to offer Ethernet-over-SDH capabilities.

"Marconi has to have this sort of product if it's going to retain its customer base and help those carriers offer next-generation services. Marconi's definitely fallen behind in the past few years in terms of developing products that offer packet multiplexing into an SDH timeslot," says Bennett. (Disclosure: Bennett is a former Marconi employee.)

So what type of operator is Marconi pitching to? Smith says its targets are incumbents and CLECs, though the alternative operators are the ones most aggressively pushing the Ethernet message to Europe's businesses. He cites Fibernet Group plc (London: FIB), which uses gear from Huawei Technologies Co. Ltd., as an example, though the U.K.-based CLEC is not alone (see Huawei Springs Surprises and Exponential-e: What Yipes Wasn't).

The incumbent carriers are also jumping on the Ethernet bandwagon, adds the Marconi man, though they're much more interested in the ability to support their legacy services, such as leased lines, as well as the new Ethernet services (see BT Offers Ethernet Over ATM).

— Ray Le Maistre, International Editor, Boardwatch

Scott Clavenna 12/5/2012 | 1:30:04 AM
re: Marconi Scales Down Its Edge This is a trend brewing, for certain. At Supercomm you could see products of this class from Lucent, Cisco, Alcatel, Fujitsu, Turin/Motorola, WhiteRock, ADVA, ANDA, Verilink/Larscomm, Haliplex, Mangrove Systems, Pontusys, Transmode, RBN, Luminous. Ocular used to have one, but I haven't seen it in action at Tellabs.

You've also got packet-oriented ones (Ethernet uplink, circuit em) from Overture, RAD, Axerra, Integral Access, Atrica, and certainly a few others.

IXCs are pushing hardest for these in the U.S. for now, but regional CLECs also see the value, since one near-term value is maximizing leased circuit utilization. ILECs will use these, too, however, when they head out of region, and when competition forces their hand at home.

Most of these devices come in under $7000, and are 1 or 2 RU. What differentiates them is whether they include integrated Layer 2 switching, standards-based GFP, VCAT, and LCAS, ring-based protection, network-side equipment protection, MPLS pseudowire encapsulation.

Definitely a space worth watching...

Scott
materialgirl 12/5/2012 | 1:30:03 AM
re: Marconi Scales Down Its Edge How fast is enterprise demand for Ethernet service growing? What speed? What is the use:
1)VLANs
2)VPN (is there a difference?)
3)VoIP replacement of "legacy" voice service?
4)WAN storage?
5)service consolidation
Do these boxes need L3 functionality for MPLS, security, subnetting or anything else?
Scott Clavenna 12/5/2012 | 1:30:02 AM
re: Marconi Scales Down Its Edge

The uptake for Ethernet services is quite fast at IXCs and the remaining facilities-based CLECs I've talked with this year. The average speed is getting a bit lower as the market expands, from around 100 Mbps last year to more in the range of 30-50 Mbps now.

The big driver is VOIP transport/backhaul for many large enterprises, and of course high-speed Internet access. Beyond that there is more interest in using Ethernet as a high-speed replacement/augmentation for Frame Relay and TDM private lines. Point-to-point Ethernet access to VPNs is a clear application as well, and in this case operators are using Q-in-Q in many cases, but are exploring using MPLS as these services scale.

The role of the Micro-MSPP in most cases in service consolidation, i.e., getting as many services onto a single pipe as possible. This can be accomplished in a number of ways and drives the design of the Micro-MSPP:

1) Mapping all the services into GFP and assigning them virtually concatenated channels within an OC-3/12/48 uplink. This is perhaps the most costly option, but solid as a rock and secure as fort knox.

2) converting all services into Ethernet (including TDM) and assigning them VLAN tags for backhaul to a concentration point in the metro. In this model, The services can be carried as a single aggregated Ethernet payload over SONET, or as pure Ethernet transport, depending on the legacy network. Some customer isolation concerns linger with this option, as do scalability challenges.

3) The future case, is converting everything (L1 and L2 services) into pseudowires using IETF PWE3 techniques and using MPLS for CoS, protection, etc. Underlying transport can be Ethernet or SONET/SDH, depending. This is a rather immature technology today, but very promising. Watch this space for more.

RPR can be used in all these cases as a Layer 2 traffic management layer within SONET or Ethernet for further efficiency gains and ease-of-networking in the metro.

Thanks,

Scott



How fast is enterprise demand for Ethernet service growing? What speed?
What is the use:
1)VLANs
2)VPN (is there a difference?)
3)VoIP replacement of "legacy" voice service?
4)WAN storage?
5)service consolidation
Do these boxes need L3 functionality for MPLS, security, subnetting or anything else?
particle_man 12/5/2012 | 1:30:01 AM
re: Marconi Scales Down Its Edge Scott,

I hope it's not too transparent that we are tapping you for free consulting time :-), but your last category of PWE3 is an interesting contrast. It would seem like that connection would require around 3 more layers of network beneath it (measuring service to physical) than a simple GFP box. In my mind more layers means more expense. Am I missing something?

PM
opticalmarker 12/5/2012 | 1:30:01 AM
re: Marconi Scales Down Its Edge Given this infusion in this space, what Global market size are they all looking at for the regular edge MSPP and also for it's micro cousin?
Scott Clavenna 12/5/2012 | 1:29:56 AM
re: Marconi Scales Down Its Edge Adding PWE3 support to an access CLE pizza box would certainly add some expense (maybe a couple thousand $, tops) and a bit of operational complexity, compared to some of the micro-MSPPs that are strictly "services over SONET/GFP/VCAT." But, in exchange for that cost and complexity, the operator does get a few key advantages:

1) The ability to support all packetized services (not just Ethernet) transport via a single unified encapsulation technique (PWE3). What's more, each packetized service can be assigned a particular class of service.

2) Converts packet traffic over TDM to packet transport (MPLS virtual circuits) at the earliest possible point, which gives you some statistical gain at the CLE across all packet services, not just Ethernet. Though some people argue that this should happen at the metro hub, not in the CLE, for reasons of network simplicity, operations, cost, etc.

3) Uses MPLS-based provisioning techniques, if you believe that's a good thing. This lets an operator create a services at the customer's prem and provision it end-to-end across their network, wherever the tunnel takes it, which could be through an MSPP, onto an edge router, through core routers and out the other end to another edge router or MSPP. This is an interesting thing to contemplate: that packet services can traverse transport and packet gear equally well using PWE3. The payoff here is that an operator gets a lot of flexibility in supporting packet services over their whole network, not just the routers and switches.

4) Service-specific status and alarm handling for each of the services carried as pseudo-wires


So, if you believe MPLS is no longer just a traffic engineering tool for the core of IP networks, but a services tool, then it starts to make sense to see how far you can push it out in the network. The PWE3 pizza box is about as far as you can go.

Scott
materialgirl 12/5/2012 | 1:29:51 AM
re: Marconi Scales Down Its Edge The market size is either huge or meaningless. Some things you can't count. If it takes off, you need to be there, period. I have never seen a parket projection number that was within 10x of being accurate. So relax, there is no answer. Precision without accuracy is a trap.
Liberty Valence 12/5/2012 | 1:29:28 AM
re: Marconi Scales Down Its Edge market price is about $1000...
HOME
SIGN IN
SEARCH
CLOSE
MORE
CLOSE