x
Comms chips

Chip Makers Face Awareness Problem

Telecom equipment manufacturers are failing to shop around for the best deals for communications chips, judging by the results of the latest market perception survey from Heavy Reading, Light Reading's market research division.

The 2004 Communications Chips Market Perception Study indicates that most purchasers of communications chips are aware of less than half of the players in any given product category. In other words, there's a good chance that they're missing out on better products, lower prices, or better service and support. Likewise, comm chip companies are missing out on capturing new customers.

This isn't a case of Heavy Reading surveying nincompoops. The 449 folk who took the survey were screened to make sure they really worked for system vendors, original equipment manufacturers, or system integrators. Then they were given a list of 31 product categories and asked to select the ones with which they were familiar. For each of these product categories respondents were then shown a list of suppliers and asked to identify the ones they recognized (see Chips on Their Shoulders).

In all but one product category, field programmable gate arrays, respondents on average failed to recognize more than half of the players.

In some cases, the recognition ratio was much lower. On average, respondents could recognize fewer than one in four vendors among manufacturers of content processors and circuit emulation chips. And they could recognize fewer than one in three vendors among suppliers of Ethernet MAC chips, VOIP chips, modulator driver chips, pre- and post-amplifier chips, backplane transceiver chips, security processors, ATM switch fabrics, ATM SAR chips, and ATM interworking chips (see table below).

It's worth pointing out that some big names were included among players that weren't widely recognized. The most striking example of this is Intel Corp. (Nasdaq: INTC). It has products in 16 of the categories surveyed, but only made it into the top three places in terms of name recognition in eight of them (see Survey Rates Chip Suppliers).

Conversely, some startups made it into the top three spots in a handful of product categories. These include Passave Inc. (PON chips); Big Bear Networks, and Phyworks Ltd. (dispersion compensation chips); EZchip Technologies, Azanda Network Devices, and ZettaCom Inc. (standalone traffic manager chips); NetLogic Microsystems Inc. (search engines) and Cavium Networks Inc. (security processors).

Table 1: Recognition Rates
Average Percentage of Vendors Recognized by Survey Respondents Number of Vendors Listed in Product Category Number of Respondents
Telephony (PDH) Chips 42% 14 77
DSL Chips 47% 13 82
PON Chips 43% 4 48
Ethernet PHY Chips 33% 17 114
Ethernet MAC Chips 32% 12 109
Ethernet Controller Chips 40% 7 93
Ethernet Switch Chips 40% 10 106
VOIP Chips 28% 19 67
Communication Processors 38% 10 109
Laser Driver Chips 35% 11 57
Modulator Driver Chips 30% 10 35
Pre- and Post-Amplifier Chips 32% 10 40
Crosspoint Switch Fabrics 35% 8 55
Sonet/SDH Data Transceivers 27% 20 84
Dispersion Compensation Chips 36% 6 28
Framer/Mapper Chips 40% 13 84
Digital Wrapper/FEC Devices 36% 6 42
Backplane Transceiver Chips 27% 21 51
Circuit Switch Fabrics 42% 7 58
Network Processors 38% 12 117
Standalone Traffic Manager Chips 34% 12 39
Search Engines 33% 10 24
Content Processors 21% 5 29
Security Processors 31% 10 40
Packet Switch Fabrics 35% 13 57
ATM Switch Fabrics 30% 9 41
ATM SAR Chips 28% 14 46
ATM Interworking Chips 29% 8 31
Circuit Emulation Chips 24% 5 23
Control-Plane Processors 35% 11 50
Field-Programmable Gate Arrays 57% 7 105
Source: Heavy Reading's 2004 Communications Chips Market Perception Study


— Peter Heywood, Founding Editor, Light Reading


Heavy Reading's 99-page 2004 Communications Chips Market Perception Study costs $3,750 - which includes access to a searchable online database for slicing and dicing results by demographic, geography, and organization type. For more details, click here.

eschrob 12/5/2012 | 2:36:11 AM
re: Chip Makers Face Awareness Problem Waverunner,
Can you provide more info on your last comment?

If your in this industry read "Engaging the Design Flow, Design Winning Sales."

Is this a book, article, something publically available?
waverunner 12/5/2012 | 2:39:06 AM
re: Chip Makers Face Awareness Problem Dedicated FAE's on a tier one's site is the holy grail of many chip makers.

"One important observation, though, is "seller proprietary", in short, increased manipulation of your solution to interface with the greater design leads to seller proprietary which in turn leads to partnership and vendor incumbency. Vendor incumbency is a significant achievement that provides both tactical and strategic advantages over your competitors; it allows the incumbent to have a direct influence on both the strategic and tactical migratory path of the technology and user product, and unavoidably future design wins and business engagement."
--Engaging the Design Flow, Design Winning Sales

W
optoslob 12/5/2012 | 2:39:12 AM
re: Chip Makers Face Awareness Problem Brand / Product Awareness is certainly an important issue for a chip maker BUT no where near as important as winning the top three sockets. Generally you find that the top three in the market drive at least 80% of the volume while the other 20 players combined amount to 20% of total sales for a given semiconductor product. So to win you need to be focused on winning the top sockets! Often this means integrating specific custom features that are not available to the rest of the market. Usually it just requires a dedicated FAE in place at these customers to make sure that their problems get resolved that day. It is amazing how loyal customers are when they know that they have a direct hook into the design team rather than some EE drop out asking them if they have applied the power.

Given the choice I'll take a top three socket over all the brand awareness in the world.

optoslob
Peter Heywood 12/5/2012 | 2:39:15 AM
re: Chip Makers Face Awareness Problem Just to be contrarian (and also because fw23 has said such nice things about our market perception surveys and I owe him/her one), I'd like reverse the impression I gave a couple of months ago, when I was resisting the idea of doing a market perception survey of OPTICAL COMPONENTS.

At the time, I'd come around to the view that it wouldn't tell us much, partly because the potential market was small and the number of big players is quite limited.

I've changed my mind. We are in the process of putting together a Who Makes What on optical components which we hope to post in the coming weeks. We'll then go ahead with a market perception survey.

I've realized the survey could shed some light on some key issues about optical component vendors - such as the emergence of Chinese vendors and the progress of Avanex and Bookham in accumulating JDS like portfolios of products.

So, watch this space! I haven't lost faith in optical components.
Peter Heywood 12/5/2012 | 2:39:15 AM
re: Chip Makers Face Awareness Problem PS on the Who Makes What report on Electronic Chips:

http://www.lightreading.com/do...

It's still worth telling us about corrections to the lists because we will continue to update them to reflect acquisitions, bankruptcies, new startups etc.

We will continue to update the report, and we'll go back and re-use the lists for further market perception surveys in the long run.
Peter Heywood 12/5/2012 | 2:39:15 AM
re: Chip Makers Face Awareness Problem Dune is in the report.

With these market perception studies, we start off by creating a "Who Makes What" report that divides the market into product categories and then lists vendors in each category. We post this on LR and invite folk to suggest additions, deletions changes, and we constantly update the report to reflect this. When the flow of corrections comes to a stop, we use the taxonomy and vendor lists as the basis for a market perception survey.

Bottom lines:

(a) We had a very comprehensive list of vendors to start with.
(b) You can easily see which vendors are included in the survey by looking at the relevant page of the Who Makes What on Electronic chips.

Dune is on the packet switch fabric page, which can be found here:

http://www.lightreading.com/do...

If you want to look at lists of vendors in other product categories, a hyperlinked list of product categories can be found here:

http://www.lightreading.com/do...



dishwasher 12/5/2012 | 2:39:16 AM
re: Chip Makers Face Awareness Problem Is there anything known on Dune Networks' (http://www.dunenetworks.com/) current products or plans? Are they in the report?
I believe they ambitiously design their NP with a 10 year life cycle and ultimate scalability in mind. At least on paper...

From some older PR:

Any port rate can be interconnected through an FE200-based switching system. Interoperability of different rates through the FE200 fabric is guaranteed. A system vendor can start with 10Gbps port rates and scale to 40Gbps and higher future rates of 100Gbps and 160GBps, all switched via the same FE200 device.

Any service scheme is supported via the FE200 devices. For example, a system can start with Ethernet/IP/MPLS service line cards, providing weight and rate-based services with tens of thousands of queues per port, and later add ATM line cards, providing finer-grained scheduling, hierarchal scheduling and millions of queues per port if required, all interconnected via the same FE200 switching devices.

The FE200 fabric is a self-healing fabric. 1+1, N+1, and N-1 fault-tolerance schemes are supported. The FE200 fabric supports a graceful performance degradation scheme, guaranteeing interconnectivity regardless of functioning fabric capacity. No software intervention is required to detect and isolate faults. Furthermore, for the first time, the fabric supports loss-less recovery from scheduled and un-scheduled faults. This means that, an active fabric card can be pulled from the system (scheduled or unscheduled) without losing even a single cell.
The FE200 switching device enables a wide range of system designs, from 80Gbps to 80Tbps, packaged in a pizza-box, integrated chassis, or multi-chassis configuration.
bkumar1 12/5/2012 | 2:39:17 AM
re: Chip Makers Face Awareness Problem Yes. It may because

There is new technology coming which is in between ASIC and FPGA.. called structural ASIC(SA). which might take things away from traditional ASIC and FPGAs.

SAs have flexibility of FPGAs like ECOs can be done any time yet with minimum NRE costs(but slightly higher than ASIC).

check-out this EE article
http://www.eedesign.com/column...



fw23 12/5/2012 | 2:39:32 AM
re: Chip Makers Face Awareness Problem Great idea!

I would also suggest getting involved in pay-for
-panel ads on search engines like google. That
way, when people go looking for a certain type
of part, they will end up right at your door.

Its an easy way to increase awareness. Even
startups with a limited budget can afford that
sort of an awareness campaign.

HEAVY READING did a great service to the
components industry by pointing out the awareness
crisis in the sector. The companies building
these electronic do-dads need to stop burying
their heads in the sand and go after the awareness
issue.

Its simple. You can't sell products unless people
know that you build that product. You can whine
about not having enough money, then you should
either have raised more money from your investors
or else not have thrown your hat in the ring
in the first place.
alchemy 12/5/2012 | 2:39:40 AM
re: Chip Makers Face Awareness Problem fw23 writes:
Another thing the guys told me is that FPGAs
turn responsible designers into hackers. They
cease to be careful and you end up with
spaghetti hardware rather than a sound design.


Modern FPGA design requires the same discipline as any other kind of hardware or software engineering. Modular design, well-defined interfaces, design for testability, detailed unit test on each subsystem... You can make an ASIC that doesn't work just as easily as you can create FPGA code that doesn't work.

If I'm building a product, why would I want
a limping wounded product even if I had it
faster? Whats the advantage to having something
faster that is damaged?


I take it you've never worked for an early stage startup that was living off seed money or angels? The way you pry money out of VCs is to show them good-steady progress. It's tough to do that when you have to get a multi-million dollar ASIC project built first.

I know that procket, which is bleeding edge in
everything, took a look at this stuff and
decided that total custom electronics was the
way to go.


Most startups are clever integrators. They leave the ASIC work to the chip companies. Personally, I try to never be an early adopter for anything and I want to see a prototyping platform/reference design and a reference from at least one successful design-in before I'll select a component. When the chip vendors are showing me PowerPointWare and telling me that the part tapes out in another month, I scratch them from my list.
technonerd 12/5/2012 | 2:39:41 AM
re: Chip Makers Face Awareness Problem I mean lots of the ===MOVERS SHAKERS===
in this crazy business concluded last year
that electrical networks are on the downhill
slope to oblivion. LIGHT based communications
is poised to rapidly displace all of this
electrical stuff.

Ain't it great to see the speculators back in the pool?
PO 12/5/2012 | 2:39:41 AM
re: Chip Makers Face Awareness Problem "I'm wondering if anyone in the peanut gallery
has any opinions on the future of FPGAs. I've
been told by several people that FPGAs were a fad that came about because many companies could not afford to do real designs anymore and because of the shortage of brain-boys who build these things.

...

"Another thing the guys told me is that FPGAs
turn responsible designers into hackers. They
cease to be careful and you end up with
spaghetti hardware rather than a sound design.
"

Now that much of hardware design has the characteristics of software design, the discipline of software development must be applied. You'll find plenty of software hacks, and I'm sure plenty of hardware hacks too. But you can also tell the difference between a designer and a programmer.

At the same time, software best-practises are no longer based on flowcharts, bubble diagrams, and the like: the state of the art has advanced.

We will not regress into ASIC development. We might advance into other special-purpose processors (e.g. Network Processors), but forklift upgrades are not acceptable in the market anymore.
waverunner 12/5/2012 | 2:39:43 AM
re: Chip Makers Face Awareness Problem I can see two reasons for this:

1) The survey was sent to the wrong target respondents, namely purchasing. Purchasing has little if any influence in the selection of components, most of the legwork and decison process is in the hands of the designer with purchasing having to deal with it after the fact. This group would be aware of only what's on legacy BOM's, meaning the short list of incumbent suppliers.

2)The electronic component market is made of "Gathers" in the sales force. Traditionally, they have trusted rep companies to represent their products, there are numerous reasons for which I will not mention here, these reps are nothing more than gathers of information, they do little in promoting the products they represent, and you know what I would do the same because they are levelled with 30day contracts, and are dropped and added depending on which way the wind is blowing.

Same goes for the internal salesforce, ask them about their accounts and they could tell you applications, design wins, the designer's name, the project name and that's where it ends, passive gathering of information. They do not active tacticians and do little in engaging revenue.

If your in this industry read "Engaging the Design Flow, Design Winning Sales."

Waverunner
opticalwatcher 12/5/2012 | 2:39:43 AM
re: Chip Makers Face Awareness Problem "If I'm building a product, why would I want
a limping wounded product even if I had it
faster?"
fw23, read!
Here's the post you replied to:
"You can also get your product limping on partially complete designs and implementations....and that can be the difference between getting VC funding or not." Also, FPGAs are good for prototyping. Even Procket did some 'partially complete' prototyping of their software using PCs.

Certainly if you are designing your own bleeding edge network processor or switch, like Procket did, then full custom ASIC is the way to go, especially if you have lots of VC money during the bubble period.

For the rest of us mortals who buy these products rather than design them, we use FPGAs to connect everything together. Maybe if the company hits it big you can convert the FPGAs to an ASIC for lower per/unit cost (but much higher NREs).

FPGAs are NOT a passing fad. Xilinx made over $1 billion last year.
icenine 12/5/2012 | 2:39:43 AM
re: Chip Makers Face Awareness Problem Gaining market share and awareness for startups is no just a matter of spending money in the media. Yes, all bases need to be covered. Today the Approved Vendor List is king. Cisco, Nortel, Juniper, etc. have AVLs that freeze out the startups.

Getting a meeting at one of these big guys to explain your new technology is easy. They love to learn what is going on. Turning that to a design win that makes the big guy dependent on the start up? Not so likely.
alchemy 12/5/2012 | 2:39:44 AM
re: Chip Makers Face Awareness Problem fw23 writes:
I'm wondering if anyone in the peanut gallery
has any opinions on the future of FPGAs. I've
been told by several people that FPGAs were
a fad that came about because many companies
could not afford to do real designs anymore
and because of the shortage of brain-boys
who build these things.


In startup-land, FPGAs are incredibly valuable. You can get a product to market quickly without the huge expense of going the ASIC route. You can also get your product limping on partially complete designs and implementations. You'd never dream of doing that with an ASIC and that can be the difference between getting VC funding or not. You also have the ability after-the-fact to fix bugs. As complexity increases, so does the likelyhood that your test process won't catch all the bugs. Personally, if an engineer walked into my office at my startup company and said we should build an ASIC, I'd thow him out the window.

fw23 12/5/2012 | 2:39:44 AM
re: Chip Makers Face Awareness Problem
Another thing the guys told me is that FPGAs
turn responsible designers into hackers. They
cease to be careful and you end up with
spaghetti hardware rather than a sound design.

If I'm building a product, why would I want
a limping wounded product even if I had it
faster? Whats the advantage to having something
faster that is damaged?

I know that procket, which is bleeding edge in
everything, took a look at this stuff and
decided that total custom electronics was the
way to go. That pre-made chip packages,
FPGAs and such were the wrong direction. They
used to have a great paper on the logic they
used at their company website.
fw23 12/5/2012 | 2:39:45 AM
re: Chip Makers Face Awareness Problem
I'm wondering if anyone in the peanut gallery
has any opinions on the future of FPGAs. I've
been told by several people that FPGAs were
a fad that came about because many companies
could not afford to do real designs anymore
and because of the shortage of brain-boys
who build these things.

I look at the survey and it looks like the
decision makers have given up on FPGAs. Like
its a dead end technology that everyone on
the shop floor is washing their hands of now
that GRADE-A talent is back available on the
market again.

fw23 12/5/2012 | 2:39:46 AM
re: Chip Makers Face Awareness Problem
Could it be that the disruptive change from
electrical to optical components in the world's
networks also partially explains the results?

I mean lots of the ===MOVERS <and> SHAKERS===
in this crazy business concluded last year
that electrical networks are on the downhill
slope to oblivion. LIGHT based communications
is poised to rapidly displace all of this
electrical stuff.

So rather than an awareness problem, maybe
its that most people at the companies looked
at by the SCIENCE of HEAVY READING just are
not up anymore on legacy technology that for
them has gone the way of the do-do bird.

I wonder in five or ten years time, as LIGHT
based communications grows, if we will look
at these "chip" things the same way we look
at Vaccume tubes today.

I would guess that Vaccume tube manufactures
might get higher awareness ratings than some
of the dead-enders in electronics like Easy
Chip.

</and>
fw23 12/5/2012 | 2:39:46 AM
re: Chip Makers Face Awareness Problem
More science from HEAVY READING. When I
look at the -->TOP LINE<---, in all this, I
see a whole industry with problems similar to
troubled Juniper networks. Lots of brain-boys
pouring cash into techno-bees down on the plant
floor while AWARENESS is IGNORED.

I was really shocked at how badly a formerly
/*\*/RED--HOT/*\*/ company like Intel had
STUMBLED in brand awareness. Too many trips
up to Redham for those boys I think.

The -->ACTION PLAN is obvious for those who
are actually LISTENING rather than having
their h-e-a-d-s i-n t-h-e s-a-n-d.

To make money, people have to know what you
make. Brand awareness and marketing are the
[KEY] that fits that lock.

Its too late for the superbowl, but a word
for those in the know ---FRIENDS--- HINT! HINT!
in the SPRINGTIME are not a bad think to have.

And for a company like Intel, Stars on america's
mind like Ray Romano and Matt Leblank (the new
Tony Danza) would be great spokespeople for those
with the BUCKS.

For startups who think they dont have the money
to play the media game, I can offer the following
pearl of wisdom to the peanut gallery....
....get big or get out!

HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE