x
Optical/IP

Marconi's First Softswitch Sale

Marconi plc (Nasdaq/London: MONI) announced its first customer win for the SoftSwitch XCD5000 (see Marconi Sells SoftSwitch).

Jersey Telecom, the incumbent carrier on the British island of Jersey, says it will use the gear to introduce broadband voice and multimedia services across its network. Details of the contract were not disclosed, but a Marconi spokesperson admits that this is not a significant win in terms of the dollar amount. But it is a strategic win for the company.

Why? Marconi announced this product back in June of 2001 at the Supercomm tradeshow in Atlanta, and to date it has not announced any customers for it. Although the company claims it's in several trials right now, any named customer win at this point is worth noting.

The island of Jersey, which sits in the English Channel between England and France, is only 9 miles long and 5 miles wide -- smaller than Manhattan with only a fraction of the population. It’s home to roughly 87,000 residents and 5,000 businesses, most of which are involved with off-shore finance. Almost every national and international bank has a presence in Jersey. About £1 billion is invested in the island’s banks at any one time, says a spokesperson for Jersey Telecom.

As a result, Jersey Telecom has become a sophisticated adopter of new technology. It has already been offering a corporate voice-over-IP solution to some of its enterprise customers. In order to offer this service, it resells smaller PBX-replacement softswitches from Nortel Networks Corp. (NYSE/Toronto: NT) and Mitel Networks to enterprise customers. Now it is looking to extend VOIP services across its entire network. The long-term plan is to replace Marconi’s System-X, a digital public telephone system, with new carrier-class softswitches, says the spokesperson. System-X currently links Jersey Telecom’s fixed telephone and data infrastructure to the rest of the world.

From a technical perspective, this is an important deployment, but because of Jersey’s size it is not likely to ever be a significant financial win. Marconi will need to find larger, incumbent carriers to also adopt its softswitch. This could be difficult, given the current carrier climate (see Softswitches Head for a Shakeout).

For example, Sonus Networks Inc. (Nasdaq: SONS), the only publicly held company dedicated to softswitches, forecasts a 60 percent revenue slide for its third quarter. The company’s revenues were down roughly 60 percent in the second quarter, too. Lucent Technologies Inc. (NYSE: LU), which was a pioneer in early softswitches, has decided to stall any further development until the market matures a bit more (see Lucent Clarifies Product Strategy).

Other key players in the market include Nortel, Cisco Systems Inc. (Nasdaq: CSCO), Siemens AG (NYSE: SI; Frankfurt: SIE), and Telica Inc.

Nortel has won deals with large carriers in the U.S. such as Sprint Corp. (NYSE: FON), Qwest Communications International Inc. (NYSE: Q), and Verizon Communications Inc. (NYSE: VZ) for its Succession Communication servers. The company has also won several European and Asian deals (see Tiscali, Nortel Deploy VOIP). Siemens is also making a big push with softswitches (see Why Siemens Sold Unisphere).

— Marguerite Reardon, Senior Editor, Light Reading
www.lightreading.com
Packet Man 12/4/2012 | 9:09:11 PM
re: Marconi's First Softswitch Sale Your comment:
But a lot of pure IP folks tend to look at an ATM network and say "hey, I can't run IP services over this without buying routers, so why do I need those ATM switches?"

My comment:
I think that to be very true, I guess cause I'm one of those IP folk. :o)

Your comment:
2. What if you already have an ATM network, you paid for it last year, it runs just fine, and is nowhere near full capacity?

My comment:
Then I would stick with the ATM, let it do its job which I'm sure does well. Extract maximum ROI from it. As time moves on, then put in plan to add IP as needed. Maybe move ATM out to the edge.

Your comment:
3. I have ATM coming out of the back of every DSLAM that I'd consider putting into the network today, so how else would I support by ADSL infrastructure?

My comment:
I agree with that too. If you got DSLAMs everywhere and the ATM network to support it, then same answer as above, extract maximum ROI from it. Later push ATM out to the edge to support the DSLAMs and use IP in the core.

Your comment:
Converged infrastructures, be they packet or cell-based, connectionless or connection oriented, routed, signalled or provisioned, *will* be less expensive to buy and run than a bunch of service-specific networks.
- It's also clear that whatever converged transport network you choose ought to be the same architecture and addressing scheme as the services it supports. It's at this point that things get more contentious :-)

My comment:
I agree.

Your comment:
- If you think that IP is the future, then clearly an IP infrastructure, either connectionless like Sprint's vision, or over MPLS like most other folks plans, is the ideal.

My comment:
This is pretty much what I think.

Your comments:
But are IP services actually profitable?
- For a multi-service carrier (ie. one that has lots of voice, general data and leased line services), IP isn't the obvious choice. It has no end to end QoS capability, and all of the services that make money will need some form of adaptation layer. I may as well use ATM adaptation layers as anything else because I know they work.
My comment:
I believe mass IP deployment can be very profitable, and I believe the way to do it is to eliminate as much of the "other stuff" (See "Old Stuff" below) as possible. If you have new stuff then I respect getting maximum ROI out it, but if its old and been paid for a 100 times over then lets start getting that stuff out. You're wrong when you say IP has no end-to-end QoS. Well, no, you are right, its not end-to-end QoS, but a well engineered IP network can deliver excellent QoS end-to-end, it would just be done on a per hop basis. With respect to the services that make money needing an adaption layer, I'm not sure I follow you? What service can we sell a customer that cannot be deliver over IP today?

My "Old Stuff" list:
Circuit-switches (Class3/4/5 DMS, 5Ess, etc), copper-loop telephones and modems and the dial-up modem banks, SS7, copper-loop based security alarm systems, closed-circuit video, TDM based TV distrubution systems, X.25 services, etc.

I could make that list alot longer. I know we can do it all over IP tonight, but we can do it over say 5 years. Once done the savings would be worth it. Just think of all the variety of staff you need to run all these different technologies, and the training, etc,etc,etc to maintain this staff. There alone would be a good place to start. I worked at a CLEC and I sized up the place one day. We had over 48 racks of equipment. I did some thinking and researching and I got that C.O. down to 9 racks, and all very do-able over IP. Ripping out that poopy brown DMS was the biggest space/power savings.


PM
hyperunner 12/4/2012 | 9:09:40 PM
re: Marconi's First Softswitch Sale Packetman,

It's extremely difficult to make apples-for-apples comparisons, especially when the data I need to back it up is covered by confidentiality agreements :-) Besides which, this data only covers one carrier, and we're all different!

Broadbandboy is right on the money with his last couple of posts. But a lot of pure IP folks tend to look at an ATM network and say "hey, I can't run IP services over this without buying routers, so why do I need those ATM switches?"

Two issues arise from this approach when it's applied to (many, not all) real customers:

1. What if you don't really think that IP is the centre of your universe right now? More on this below.

2. What if you already have an ATM network, you paid for it last year, it runs just fine, and is nowhere near full capacity?

3. I have ATM coming out of the back of every DSLAM that I'd consider putting into the network today, so how else would I support by ADSL infrastructure?

4. What am I going to use to support my UMTS infrastructure...MPLS? LOL

But now some general points, all of which are my own opinions/experience, not "facts".

- Converged infrastructures, be they packet or cell-based, connectionless or connection oriented, routed, signalled or provisioned, *will* be less expensive to buy and run than a bunch of service-specific networks.

- It's also clear that whatever converged transport network you choose ought to be the same architecture and addressing scheme as the services it supports. It's at this point that things get more contentious :-)

- If you think that IP is the future, then clearly an IP infrastructure, either connectionless like Sprint's vision, or over MPLS like most other folks plans, is the ideal. But are IP services actually profitable? A lot of the OpEx savings that IP promises are based on the assumption that you'll run automated provisioning and constraint-based routing (if you don't assume this you'll be stuck with manually provisioned TE, which is no better than ATM today). Many carriers don't trust the idea of dynamic routing protocols, and won't even turn PNNI on in their ATM switches, so I'm pretty sure thay'll be wary of OSFP-TE and RSVP-TE in a CBR scenario.

- For a multi-service carrier (ie. one that has lots of voice, general data and leased line services), IP isn't the obvious choice. It has no end to end QoS capability, and all of the services that make money will need some form of adaptation layer. I may as well use ATM adaptation layers as anything else because I know they work.

- But what about new IP services, like VPNs? It's clear they're profitable, so that must be good eh? Not necessarily. If my new IP VPN services cannibalise revenue from my existing Frame Relay, SMDS, ATM and leased line services, I'd be pretty stupid to invest my own money to assist that cannibalisation, wouldn't I?

- Aha, but what about competition? If I don't offer these new services, one of my competitors will. Again, that's not obvious these days. Most of the carriers who are still solvent (technically, anyway) are in the same position. There's a real dilemma over how fast and easy to make the rollout of the new services. And when these services start up, it's actually easier and cheaper (ie. lower cost of sales and marketing) for the sales force to target your existing business cutomers (even the dumbest carrier knows where at least 80% of its customers live ;-)

There are a lot of things that could change this situation...

- If a bunch of carriers weren't saddled with debt (WCOM maybe? Scary)...

- If voice demand wasn't flat or falling...

- If there really were a compelling application for business broadband...

- If there really were a compelling application for residential broadband...

- If IP services generally were A LOT more profitable...

- If IP VPNs really could offer QoS guarantees (instead of only guaranteeing priority between the CE and PE routers)...

- If OSPF and especially BGP were much more stable and linked with a bigger, structured address space (like ATM Private Addresses and PNNI, for example)...

- If carriers didn't think they needed to buy two routers for resilience...

- If only MPLS had been a real merging of the best of IP with the best of ATM, instead of turning into an inter-vendor battleground...

- If only Microsoft would announce that IPv6 will be the default network protocol for the 2005 release of Windows :-) ...

...we could call up all our buddies who are out of work and give them the best Christmas gift they could wish for :-)

hR.
Packet Man 12/4/2012 | 9:09:49 PM
re: Marconi's First Softswitch Sale You wrote:
For example, several years ago I was told by the head network engineer for a large ISP that one guy could manage five ATM switches, but he needed to assign one BGP engineer per router. And skilled BGP guys are not cheap! I suspect that is what hyper is talking about.

My comment:
Hmmmm never looked at it from a BGP point of view. But if an ATM switch were stuffed up with IP routing logic and it had to hold the Internet BGP table would it not be as equally costly??

Your comment:
As for Telus, they do say they plan to run voice over IP and get rid of ATM, but the fact is, they are installing a VoATM system on the edge. That fact is not mentioned in the article, so like I said, don't believe everything you read!

My comment:
I'd do the same, push my legacy core technology out to the edge. Someday GSR-IP will be pushed to the edge and Photonic-IP-MPLS-wingnut will be in the core. However with respect to Network World article it is quoted "Pathak said Telus would eliminate the ATM layer in its network, taking care to ensure ATM users are ported onto IP without problems." That does not sound like ATM will be at the edge. Personally I trust propellors heads like me and you over the marketing/sales types.

Me
broadbandboy 12/4/2012 | 9:09:51 PM
re: Marconi's First Softswitch Sale Packet Man wrote: "First, I have never read an article that stated IP costs (capitol and/or operational) were higer than ATM costs. I have however read numerous articles that one of ATMs few
pitfalls is its higher costs (both capitol and operational) than other technologies."

Packet Man, you have been misinformed. A word of advice from someone who knows, most of what you read in the press is just plain wrong.

Hyperunner is speaking from experience when he said that ATM was less expensive to operate than IP. So this is not a cost comparison of hardware, but the all important operational costs which are much more significant.

For example, several years ago I was told by the head network engineer for a large ISP that one guy could manage five ATM switches, but he needed to assign one BGP engineer per router. And skilled BGP guys are not cheap! I suspect that is what hyper is talking about.

As for hardware costs, I don't have specific list prices for a GSR 12400, but I don't think they are that much cheaper than a typical ATM core switch. Maybe someone else can fill us in on the details.

Your point that, all things being equal, IP over Sonet is cheaper than IP over ATM over Sonet, is of course correct. But it would not be so cheap if routing instability were every to take down a major ILEC voice network. So there may be some good reasons to pay a little more for layer-2 reliability.

As for Telus, they do say they plan to run voice over IP and get rid of ATM, but the fact is, they are installing a VoATM system on the edge. That fact is not mentioned in the article, so like I said, don't believe everything you read!

Cheers,

BBboy



Packet Man 12/4/2012 | 9:09:52 PM
re: Marconi's First Softswitch Sale Second is the perception of IP. Many carriers have been burned by the high operational costs of their IP networks, and softswitch networks are often thrown in the same pile.

However, if you're talking Voice over ATM, I'd fully agree that the business case is solid. As I've said a number of times on these boards, I believe that ATM switches cost much less to run than a comparable router network (no specific statistical evidence, just personal experience). In addition, VoATM softswitches have the same, or better operational savings as VoIP softswitches. Note, I say they may have better savings because VoATM switches and gateways were all designed for carrier operation, whereas VoIP stuff may be scaled-up enterprise gear.
------------

I disagree with that comment.

First, I have never read an article that stated IP costs (capitol and/or operational) were higer than ATM costs. I have however read numerous articles that one of ATMs few pitfalls is its higher costs (both capitol and operational) than other technologies.

Secondly lets assume that IP technology was twice as much as IP technology. Lets use simple math. ATM = $100, and IP = $200. To deliever services one has three choices, Pure ATM, Pure IP, or IPoATM. Since we had the Internet we have no coise but to deploy IP. We cannot eliminate IP from our network, unless we are a strict "raw circuit" carrier. So for most of us its the IPoATM model. Simple math gives us $300. However 99% of the services are now deliverable via IP or an MPLS-IP network. Therefore we can elimaite ATM completely. Our simply math now gives us $200. So our 100% converged IP network using this simply math is 66% of the cost of todays network, even if we assume IP is twice as much cost as ATM.

Thirdly, some softswitches (agreed, not all) are not enterprise boxes that have been stuffed up. Some vendors have made a box from scratch with the day 1 goal of being 'carrier class'. Those boxes can do ATM, IP, IPoATM, etc. Therefore their VoATM and VoIP costs are comparable. However as I said in my first paragraph I have always read that ATM technology is more expensive than IP technology. Secondly an IPoATM network is always more than an 100% IP network. Get rid of the ATM !!!

A Canadian carrier as announced it is no longer going to invest in ATM and will also be replacing its ATM with a 100% IP network, to help it reduce its costs. See URLs;
http://www.itbusiness.ca/index...
http://www.itworld.ca/portals/...

Overseas, Norways Lyse Tele Builds Ethernet FTTH with Cisco Metro Network. No ATM there either. Just IP/802.3/Ethernet/Fiber. I used to have the URL for it but I can't find it. Here is my post regarding that:
http://www.lightreading.com/bo...

Me
hyperunner 12/4/2012 | 9:10:03 PM
re: Marconi's First Softswitch Sale broadbandboy,

I agree with you 100% about Sprint, but they seem to be in the minority at the moment - namely carriers who are willing to spend some money in order to reduce OpEx.

If you ask any carrier they'd say "of course we'll spend money if the business case justifies it". But in reality there are two things that seem to stop that happening.

First is that a lot of them are just talking the talk.

Second is the perception of IP. Many carriers have been burned by the high operational costs of their IP networks, and softswitch networks are often thrown in the same pile.

However, if you're talking Voice over ATM, I'd fully agree that the business case is solid. As I've said a number of times on these boards, I believe that ATM switches cost much less to run than a comparable router network (no specific statistical evidence, just personal experience). In addition, VoATM softswitches have the same, or better operational savings as VoIP softswitches. Note, I say they may have better savings because VoATM switches and gateways were all designed for carrier operation, whereas VoIP stuff may be scaled-up enterprise gear.

hR.
broadbandboy 12/4/2012 | 9:10:07 PM
re: Marconi's First Softswitch Sale hyperunner: "IMHO VoIP in the carrier network is stalled because the current generation of circuit switches do the job just fine, and it'd cost a fortune to rip out the old infrastructure and replace it with packet voice products. Given that there's liitle or no growth in demand, we can't rely on new builds with packet switches to gradually replace the installed circuit switching base."

hyper, your point about no growth in demand is a good one, because just a few years ago, circuit exhaust was a big problem and the choice was between buying new circuit switches or new softswitches. That has pretty much gone away except on the access tandems.

Nevertheless, I believe that the ILECs still need to reduce their cost of operations. Based on what has been written about Sprint, they will replace 300 circuit switches with about 80 ATM switches. Somebody correct me if I am wrong, but 80 ATM switches has got to be a lot cheaper to deploy, operate and maintain than 300 old-style circuit switches. The result is more money going to the bottom line. In fact, the operational cost savings should pay for the costs of transition, and then some.

BBboy

broadbandboy 12/4/2012 | 9:10:08 PM
re: Marconi's First Softswitch Sale Raypeso: "The company I work for has purchased several softswitches and is very committed to moving into VoIp and retiring the old Class 5 switches. The new softswitches have been in our network for quite a while and still don't carry any customer traffic, main reason being that they are unreliable.
Customers who are used to dependable service from Lucent and Nortel switches will find it difficult to accept all the problems of the new breed."

-------------------------------------

Hi Ray,

Are your legacy class 5s are Nortel DMS or Lucent 5Es? Nortel has installed its Succession softswitch gear at several of its DMS customers. They are operational at Verizon. Sprint and Telus are planning to deploy. It seems that by sticking with the incumbent switch vendor, they don't have as many problems with integration and feature compatibility.

One would assume that this product has been pretty well shaken out if the ILECs are deploying it. I have also heard that Santera has a platform that actually works in live networks, but I don't have first hand knowledge.

BBboy
stbuk 12/4/2012 | 9:10:27 PM
re: Marconi's First Softswitch Sale The Hon Raymond G H Seitz is actually American - ex-US ambassador to the UK.

I think the US does use "Honorable" for certain positions, judges, others?
hyperunner 12/4/2012 | 9:10:27 PM
re: Marconi's First Softswitch Sale A while back we had presentations on Marconi's softswitch, and one of the big features they talked about was its ability to deal with videoconferencing streams.

IMHO VoIP in the carrier network is stalled because the current generation of circuit switches do the job just fine, and it'd cost a fortune to rip out the old infrastructure and replace it with packet voice products. Given that there's liitle or no growth in demand, we can't rely on new builds with packet switches to gradually replace the installed circuit switching base.

But I'm seeing a lot of demand by large enterprises for packet based video conferencing. I know a lot of trading houses, financial analysts, etc. in the City of London have tossed out the old ISDN VC gear and installed IP devices recently. Payback is within weeks.

But, like current VoIP deployments, these are private video networks. Even so there's a demand by companies to be able to link to their customers and supply chain partners, and that's where a carrier could step in to offer a multi-media service over, say, a VPN service.

That service would also be popular with smaller businesses who can't afford to build their own, private video WAN.

What do you folks think?

hR.
hyperunner 12/4/2012 | 9:10:27 PM
re: Marconi's First Softswitch Sale Beachboy wrote:
Shock, horror ! A venerable British institution replacing Sirs, Lords and Honorables.


I'm with you man. But I asked my Brit buddies here and they were pretty unanimous that these outgoing dinosaurs will just be replaced by another bunch of "Sir", "Lord" and "Your royal highness" dead wood in the new board.


By the way, I gather the title "Honourable" applies to UK Memebers of Parliament. Judging by the ongoing lack of honour in the two main parties here this is a pretty good joke. It's just like being back home :-)

hR.
Packet Man 12/4/2012 | 9:10:33 PM
re: Marconi's First Softswitch Sale The vendor is Sonus. As far as I know we haven't looked very hard at anyother but I could be mistaken.
---------

The two vendors I heard good success with are Taqua and Telica, with Taqua being a Class 5, and the Telica (Plexus 9000 I think) is a Class 4 and 5 in the one box.

Anyone else have any opinions about these three vendors, good or bad?

Me
raypeso 12/4/2012 | 9:10:34 PM
re: Marconi's First Softswitch Sale We have problems with echo. The equipment it'self has been shaky. On a few occasions we've had card failures that have dropped the entire switch. If there had been live customer traffic on it we would have been in real trouble. Things are improving though, we began putting some of our SS7 traffic over the switches, the redundant path. The vendor is Sonus. As far as I know we haven't looked very hard at anyother but I could be mistaken.
Packet Man 12/4/2012 | 9:10:37 PM
re: Marconi's First Softswitch Sale The new softswitches have been in our network for quite a while and still don't carry any customer traffic, main reason being that they are unreliable. I have no doubt in my mind that if the product worked as well as our old switches that they would be in full use by now.
These things fail in such fantastic ways that sometimes it amazes me. Customers who are used to dependable service from Lucent and Nortel switches will find it difficult to accept all the problems of the new breed.
--------------

What kind of problems are you having? I know two companies that have not run into any major issues. Just minor issues, and I'm sure DMS/5E also come with their share of minor issues. What kind of fanstasic problems are you having? What vendor are you using?

Me
link 12/4/2012 | 9:10:38 PM
re: Marconi's First Softswitch Sale "Customers who are used to dependable service from Lucent and Nortel switches will find it difficult to accept all the problems of the new breed."

So has your company evaluated the softswitch from Nortel and in your opinion, is this in the same boat as the "new" softswitches in your lab?
raypeso 12/4/2012 | 9:10:42 PM
re: Marconi's First Softswitch Sale I think the outlook for VoIp switches would be far better if the vendors could produce a working product.
The company I work for has purchased several softswitches and is very committed to moving into VoIp and retiring the old Class 5 switches. The new softswitches have been in our network for quite a while and still don't carry any customer traffic, main reason being that they are unreliable. I have no doubt in my mind that if the product worked as well as our old switches that they would be in full use by now.
These things fail in such fantastic ways that sometimes it amazes me. Customers who are used to dependable service from Lucent and Nortel switches will find it difficult to accept all the problems of the new breed.
beachboy 12/4/2012 | 9:10:44 PM
re: Marconi's First Softswitch Sale From News Feed :

Sir Alan Rudge and Hon Raymond G H Seitz, both non-executive directors of GEC plc and Marconi plc since 1997, have announced their intentions to resign from the Board immediately to make way for an orderly reconstruction of the Board. Sir William Castell and Nigel J Stapleton, who joined the Board of GEC plc in 1997 and Marconi plc in 1999, did not offer themselves for re-election to the Board at the company's annual general meeting on October 8, this year.

Shock, horror ! A venerable British institution replacing Sirs, Lords and Honorables.
I've always felt that either the inclusion of these guys or the 'promotion' to Lord or whatever is at the expense of credibility in the 'tech' business world. It conjures up too many stuffy and dated images.
dljvjbsl 12/4/2012 | 9:10:54 PM
re: Marconi's First Softswitch Sale
The prospect of Softswitch deployment in the public carriers is rather dim. VoIP is facing similar fate except its limited usage in the enterprise ernvironment. The Softswitch about 2-3 years ago were pleading a very rosy picture of its deployment prospects. They were targeting every customer under the sun. This included ILECs, CLECs, IXCs and ISPs. Overnight over twenty companies sprang up in the US alone with the help of the VCs.


This is very true except for the comment about VoIP -- limited usage in the enterprise ernvironment. This is a very significant market which will be getting much larger and much more signficant in the coming years. Teh carrier market is the one that will fade.

Wahtr is very true though is that teh soft switch is a very dumb idea that comes alive very few years (AIN, CTI, ... )in varius guises. It always fail because it solves no real customer problems that connot be done better in different ways
BobbyMax 12/4/2012 | 9:10:55 PM
re: Marconi's First Softswitch Sale The prospect of Softswitch deployment in the public carriers is rather dim. VoIP is facing similar fate except its limited usage in the enterprise ernvironment. The Softswitch about 2-3 years ago were pleading a very rosy picture of its deployment prospects. They were targeting every customer under the sun. This included ILECs, CLECs, IXCs and ISPs. Overnight over twenty companies sprang up in the US alone with the help of the VCs.

VCs do not allow scientists and industry experts to evaluate the concepts throughly. There is so much supply of money in the US to the VC community that is hard for these guys to stay within their limits. Their ignorance and greed is of great concern.

The critics have appropriately concerned their concerns in the areas of reliability, integration, standards, and cost. None of the oftswitch companies have gained any traction in the US. Clarent, which I believe is about seven years old, does not have much to show for.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE