Next-generation access will redefine the telcos
Gazettabyte caught up with him to understand the goals of his new company, Diffraction Analysis, and why he believes next-generation access is critical for service providers.
"As soon as you, the operator, make that investment decision, it has fundamental implications as to who you are as a company"
Benoît Felten, CEO, Diffraction Analysis
Gazettabyte: There are several established market research companies addressing access. What is Diffraction Analysis offering that is unique?
BF: There are two reasons [for setting up Diffraction Analysis]. The first came to me when I was doing consultancy work for a [Yankee Group] customer. He said: “You are the only guy I know working for an established company that only covers next-generation access.” All the other guys cover broadband, with next-generation access being a sub-topic.
At that moment it coalesced something that I had been thinking about for some time: the migration from legacy to next-generation access networks is probably the single most challenging issue that established players will face, and the single biggest opportunity for challengers to grab. If you drown that [topic] among legacy [broadband] issues you might be missing the point.
The second reason, much more pragmatic, is that there are many small companies that simply cannot afford the cost of generic telecom research from established market research companies. To access research affordably, for me, that is a market opportunity.
When you say next-generation access, what do you mean?
BF: It refers to the replacement of the legacy copper network in all its incarnations – most cell towers are connected with copper today - with a fibre-rich network. Cable networks, wireline copper networks, mobile networks are all going to be fibre-rich.
What are the key issues facing operators regarding next-generation access?
BF: The first for the operators is: How do we finance a network deployment and why do we do it? The established players all agree that they have to do it, sooner or later and probably sooner, and the core question is: How do we do it?
The problem is that it places access at the core of the telco business model. Ever since the internet started being successful, most legacy players – and that includes cable players - have seen themselves as service providers rather than access providers. Effectively, they are faced with a major investment which if they don’t do opens up opportunities for others to displace them. We are seeing that happen is small markets like Hong Kong, where a competitive player is on the path to eliminate the access network of the incumbent.
The threat is real, the customer need is real. The problem is operators don’t know how to use the network for their own revenues. They are faced with the choice of becoming a long-term utility – investing in the network for 20 years and reaping revenues for another 50 years – but that is unpalatable for them, or they find another way to use the network for revenues, keeping in mind that most new services do not come from telcos these days but from over-the-top players.
What we plan to examine are the alternative paths: What will be the operators’ role and where will the operators’ revenues come from once they have made this investment?
As soon as you, the operator, make that investment decision, it has fundamental implications as to who you are as a company. It is not just an upgrade.
I was at a conference last year and a guy from NTT said: “We didn’t realise that when we made that [fibre access network] investment decision, we were rebuilding the company from scratch.” He said: “Now, 10-years-on, at a strategy level, we have understood that – we are in a different business now.”
What is Diffraction Analysis going to do?
BF: We are a market research and consultancy firm. It is important to do both: consultancy keeps you grounded in what is happening in the market. Research is your ability to step back and articulate the global view.
I have already signed a couple of companies for whom I do advisory services. We also have classic consultancy projects. We are working for a vendor right now who is asking us to look at opportunities for them to enter the access market. They have disruptive technology and are looking to partner with companies and take a stake in the access market. We are in the middle of this and our advice might be: don’t do it.
One of the things we want to do is build modelling tools that allow legacy service providers to map the network deployment in time and not just based on a single investment decision. Right now the question is do I deploy fibre or not? But the reality is even if the answer is yes, the deployment will take 15 years. If it takes 15 years, what happens to all the people that don’t have fibre as I – the operator - gradually connect them?
We are trying to build a model that will optimise the cost and the service offered to end customers with a variety of technologies. This is where fibre-to-the-curb and various flavours like phantom mode DSL come into play.
We are aiming to do this by geographical area, to model where you should deploy fibre first and what you should do in non-fibre areas, and for how long, looking at the lifetime of these various technology options.*
What are the key lessons you learnt as a Yankee Group analyst?
BF: One of the things that strike me is that in this economic shift we have experienced in the last 30 years, something has been lost and that is long-term vision. That leads many organisations to make hugely inefficient decisions. These decisions may be rational but the long term is no longer part of the equation. In the telecom business it is striking how far this can lead people into making wrong decisions.
The second thing that I learnt interacting with many industry players is that the single toughest challenge each organisation has is fighting against their own culture. There is a culture of business-as-usual which is at odds with the challenges of an ever shifting technology market. Even companies in the internet space that everyone views as agile and willing to reassess themselves, you find these cultural issues.
I’m not saying anything original but interacting with these companies all around the world for the three years at Yankee highlighted this for me.
Most broadband users are still DSL-based. How will fibre-based access become massively deployed?
BF: Essentially there are three drivers for telcos to deploy. In order of importance they are: competition, network reboot and meeting customer demand.
Competition is a clear driver. When as an organisation your network access business is threatened, every consideration about how fast you deploy for payback goes out of the window - you have to deploy. And then you learn the hard way since by responding and not anticipating, you make mistakes.
The second driver [network reboot] is not mature today. Smart CTOs around the world are seeing fibre deployments as an opportunity to rethink way more than just their access infrastructure. And WDM-PON [wavelength division multiplexing – passive optical network] technology in access plays a significant part in that thinking.
If they deploy now, they may make savings and achieve network concentration but it is not massive. If they wait they might be able to save more which is why this driver isn’t working right now.
The third driver is meeting customer needs. Now, in their public discourse, operators say this is first and foremost but the reality is that since they have not found ways to make money out of traffic, they don’t want more traffic. So meeting customer needs is not a priority except if you are in a competitive market and someone else is meeting customers’ needs in which case you have to do it.
Diffraction Analysis’s team comprises people with wireline experience but the company does plan to also cover mobile. “I do think that there is a great deal of sense in having a mobile arm too but I can’t build that myself – I don’t have the credibility or the knowledge,” says Felten, who is looking at partnerships or recruitment to add mobile to the operation.
*Diffraction Analysis has just published its research programme till June 2011.
ROADMs: Set for double-digit growth
Summary
The wavelength-division multiplexing (WDM) reconfigurable optical add-drop multiplexer (ROADM) equipment market will be the fastest growing optical segment over the next few years, according to Infonetics Research. The market research firm in its ROADM Components Market Outlook report predicts that the segment will grow at a compound annual growth rate (CAGR) of 13% from 2008 to 2013.
Q&A
Q. Can you please help by defining some terms? What is the difference between a wavelength-selective switch (WSS) and a ROADM?
AS: A WSS is a component that can direct individual wavelengths among multiple fibers. They are typically built in asymmetrical configurations, such as a 1x9 or a 9x1 and are used in quantity to build logical larger switches, effectively allowing multiple wavelengths to be switched among several incoming and outgoing fibers.
ROADMs are subsystems composed of these WSS modules but also include EDFA amplifiers, splitters, sometimes arrayed waveguide gratings (AWGs), and control electronics that include power balancing.
Q. A ROADM can also be colourless and directionless. What do these terms mean?
AS: For a ROADM to be colourless, it must be capable of dropping wavelengths of the same colour entering the node from both the West and East directions on individual drop ports. Directionless requires that wavelengths added at that node have non-blocking behavior and be capable of being routed either in the West or East direction. Removing these restrictions typically requires more WSSs to be used in the ROADM in place of AWGs, representing a classic flexibility/ cost tradeoff.
Q. In the report you split the WSS into two categories: those with up to four ports and those greater than four ports. Why?
AS: That’s really the breaking point of the market according to carriers I spoke with. Originally, four ports was a high end number but since then larger WSS modules have become available. The market has divided into small, which is 1x2 to 1x4 ports, and large, which at this point are 1x9’s.
"It is probably the only thing the circuit-loving Bell-heads and the counter-culture IP-bigots can agree on – everybody loves ROADMs."
Andrew Schmitt
Q. You say that ROADMs will be the faster growing optical equipment segment. What is motivating operators to deploy?
AS: ROADMs save money, plain and simple. When you use a ROADM, you eliminate the need to do an electrical-optical conversion and the electronics required to support it. It is particularly attractive for IP over WDM configurations, where expensive layer three router ports can be bypassed. Electrical-optical conversion is where the cost is in networks and ROADMs allow any given node to only touch the traffic required at that node. It’s probably the only thing the circuit-loving bell-heads and the counter-culture IP-bigots can agree on – everybody loves ROADMs.
Q. Are there regional differences in how ROADMs are being embraced? If so, why?
AS: North America, Japan and Europe have seen the bulk of deployments. But that has started to change with smaller carriers in developing countries adopting ROADM, particularly in Asia Pacific.
Q. Did you learn anything that surprised you as part of this research?
I assembled historical estimates of the WSS market back to 2005 through conversations with WSS vendors and equipment makers. Most people were very co-operative. When I was writing the final report, I overlayed historical WSS component revenue with the Infonetics’ ROADM optical equipment revenue we have tracked over the past years, and there was an extremely tight correlation. Where there wasn’t a correlation there was a logical reason behind it – adding more ROADM degrees to existing nodes.
Covering the component market and the equipment market makes the research much better than if I did each market individually. I’ve done a lot of research in the past few years in both technical and financial domains but this was the second most interesting – it was really refreshing to find a big double-digit growth market in optical.
Cisco System’s CEO, John Chambers, has been very public in his goal to grow the company at 15% annually, and I don’t think it is an accident that the Cisco optical group makes ROADM solutions a number one priority. They’ve silently moved up to second in market share for North American WDM, and their ROADM expertise played a big role in this.
The wavelength-division multiplexing (WDM) reconfigurable optical add-drop multiplexer (ROADM) equipment market will be the fastest growing optical segment over the next few years, according to Infonetics Research.
The market research firm in its ROADM Components Market Outlook report predicts that the segment will grow at a compound annual growth rate (CAGR) of 13% from 2008 to 2013.
Andrew Schmitt, directing analyst, optical at Infonetics discusses some of the issues regarding ROADMs and his report findings.
The art of market analysis
Bob Larribeau, a telecom industry analyst and technology consultant since 1992 has just retired. gazettabyte asked him to reflect on what it takes to be a good market research analyst.
"Most companies provide good numbers but some, quite frankly, are hard to believe."
Bob Larribeau
There are several skills a good industry analyst must develop. The ability to communicate well - in writing, in presentations and in informal exchanges - is critical, as is a broad knowledge of the telecom industry and its technologies. These days the ability to have a global view - an understand of developed, developing, and emerging markets - is also important.
In addition, there are two more skills an analyst must possess.
The first is the ability to develop a detailed knowledge of the industries covered. For telecom this includes knowledge of the service providers and vendors and the products and services they offer. Regularly publishing reports that compare results and market positions of the service providers and vendors is an important part of this process.
This skill includes evaluating data provided by service providers and vendors. It is important that definitions are aligned to be able to compare results from companies. It also requires detecting when data is misleading or false. Most companies provide good numbers but some, quite frankly, are hard to believe.
An analyst may also have to develop their own market metrics since the statistics provided by the vendors are inadequate.
When I started following the IPTV market I found it necessary to develop ways to use service provider subscriber counts as a metric. This worked well for most segments, but I had to work with video encoder vendors to use the number of encoders sold. The video encoder firms were willing to provide such information to have an independent view of market position. Such metrics developed by the analyst can provide a good assessment of the market and are likely to be the only choice.
Reporting market positions is the most sensitive part of an analyst’s job. No company is happy when an analyst’s assessment of market position does not match the company's own. A few companies believe that they can improve their market position by being aggressive with an analyst. The analyst must be able to listen to the company but cannot be browbeaten into changing his or her assessment. The challenge is doing this without disrupting an important relationship.
The second analyst skill is the ability to provide a short and long term view based on the data an analyst develops. This includes identifying key trends and understanding how these trends will affect the markets they study as well as how they will affect the participants. Providing a strong strategic perspective is what characterises the best analysts.
I developed business models for several markets I studied. I found that a simple business model can provide a good perspective on the problems that service providers face. Ten years ago one of my models showed that the financial structure of unbundled access in the U.S. market would make it difficult for the competitive broadband providers to be profitable, which was indeed true.
Business case analysis also showed that WiMAX could be a strong competitor to wireline broadband and that it will be difficult for mobile TV operators to make a profit on the service.
I also developed ways of assessing future market opportunity. I showed that there were few opportunities left for vendors in the IPTV market even though there was major growth in subscribers ahead. This conclusion was based on the assumption that it will be difficult to disrupt existing service provider/ vendor relationships in most market segments. This has proved to be the case even though many aspiring vendors felt that the market would be more open.
I found being an analyst a good job. It provided plenty of opportunity for creativity and allowed me to work with a lot of great people.
Bob Larribeau has worked on his own and for RHK (now part of Ovum) where he was responsible for its access and switching and routing services. He was also affiliated with MRG where he made contributions in broadband and developed its IPTV service. In 2004 he co-founded TelecomView with Ian Cox, another RHK alumnus. TelecomView analysed the market for WiMAX as well as broadband and IPTV.
Bob’s first project was a private analysis of the commercial Internet in 1992. In that year, commercial Internet revenues were $15M. In 1999 he wrote an analysis for RHK that forecast that IP traffic was about to eclipse both ATM and circuit-switched voice traffic and predicted the coming importance of MPLS in IP networks. In 2001 he began covering the IPTV market. Bob can be contacted at bob@larribeau.com