For my father’s generation, images were hard to come by, and expensive to produce. His slides consisted of generic images … a picnic … a tiger … a train … a cartoon … designed to emphasize a point and form a memory aid, and be used over and over in different ways.  It was the speaker’s job to link the images, and build the understanding of their meaning.  For example, the picnic could represent the innocence of children with little understanding of risk, and the tiger could be an animal that they want to study.  From the feedback I’ve had since his passing, his talks and those images provided powerful links that have endured, and his work will not be forgotten for a long time to come.  (Apparently they knew he was really getting up a head of steam when his shoes came off…)

In comparison, today’s world of powerpoint and the internet provides an almost infinite supply of material, and the ability to easily customise slides in software – and we’ve got lazy. Death by Powerpoint is now a common problem, and story-telling is a dying art.  It is all too easy for slides to present facts to be read out, with a token illustration in the corner.  People read words rather than listening, and it is tempting for speakers to ‘follow the text’ in trying to make their point.  Sadly, it’s even easier to put up lots of complicated equations, which your audience won’t follow.  The result is often about as inspiring as dried mud, as the whole method is designed for information (quickly forgotten), rather than engagement (passed to memory).  There is nothing worse than watching an audience quietly start checking email en masse while the speaker rambles on. 

To be fair, audiences and speakers are now often multi-national, and using bullet points and text help overcome differences in accent and can have their place.  Complicated technical details and hard-to-pronounce jargon are also useful to set out.  But they’re not story-telling, and never going to be inspiring.  Because the points are on the page, it is tempting for speakers to do far less preparation in getting their story straight, and it shows.  The software and our laziness means that they also get double use as documents, and handouts of the actual slides are also now often used as “take-away’s”, for later study.  Trying to get a talk to do this double duty well doesn’t work – the requirements for talks and notes are different.   However, continuing advances have brought the wheel full circle, and recordings now include both sound and images.  This enables the return to story-telling, as the slides no longer need to stand alone.   

Speaking is a performing art, requiring training and practice, practice, practice. Standing on a stage with only your preparation and your wits is very confronting for most of us, and those comforting bullet points can ease the panic somewhat and guide you forward.  But they can also be a crutch that gets in the way of reaching out to your audience and truly engaging.  Whilst I hope that my talks have been a step above the average, I am guilty of bullet-point mania.  I’m resolving to take a leaf out of my father’s book, and start to use slides and simple images more creatively to provoke, entertain, engage and educate.  I think it’s a good challenge, although I’m amazed at how exposed the mere concept makes me feel.  I’m taking that as a sign that like most challenges, this one will cause growth, and bear its own rewards in time.

The Market Watch was well attended – standing room only at the back.  I took away a few key messages:

• 2009 was bad, but nowhere near as bad as the tech wreck years, and is likely to improve.  Huawei did best, driven by mobile growth in China, and are now #1 in 40G line cards
• There is a real problem with innovation in the industry – both VC and vendor investment are trending down. The VC investments in Comms that did happen in 2009 were mostly in wireless.
• Exponential growth in traffic is here to stay, and has been around long enough now that there is a clear and steady trend
• The relentless cost/bit pressure will also persist, and everything needs to get more efficient to feed this trend.
• The industry has done a lot of necessary consolidation recently.  Top 10 optical component vendors currently have 72% of market.   There is more consolidation to come, with an emphasis on vertical and cross-layer integration, and likely to be painful. 
• Telecom and datacom companies are starting to look more and more alike – causing synergies and opportunities
• 2010 is seen as ‘the year’ for massive 40G deployment:  2013 is expected to be the year of 100G
• Trends are networks moving up in bit rate, channel spacing dropping from 100 GHz to 50GHz, networks moving from rings to meshes, and ROADMs evolving from coloured to colourless and dierctionless
• The long-held trend in higher bit rate adoption of 2.5x increase in cost for 4x increase appears to be dead.  75% of carriers have said they’d be likely to adopt 100G at a 20% cost/bit improvement.  One of the drivers here is ROADMs – more channels require higher degree ROADMs to be placed in the system, while fewer fatter pipes can use the existing kit

There were also some interesting snippets for your next party conversation:

• There have now been >200 billion minutes of skype conversations
• The original development of optical fiber required 10^98 (that’s a one and 98 zero’s, or nearly a Googol) improvement in glass transparency
• Enough optical fiber has been laid in the last 40 years to do more than 4 round trips to the sun

One other interesting observation is how little twittering and on-line activity there is about the show.  Maybe the trend has taken off more in Australia than the US, or maybe the crowd here feel that Twitter is a little frivolous for their ‘serious’ industry, but the activity is noticeable by it’s absence.  There’s a LinkedIn group tag on the conference login page, but I think that’s too slow for a conference in progress.  There is no Twitter hashtag promotion, which might also have an influence.  Will be interesting to see whether this kind of activity picks up next year.

Please feel free to add your thoughts and comments on the show below

The topics were all really interesting, and I found myself torn between two or three possibilities for each timeslot.  In the end, I skipped across sessions, and captured some interesting ideas to share with you. 

Discipline-Hopping in Photonics: How Can Medical Sciences Benefit from Recent Advances in the Traditional OFC Technologies?

This workshop was based on the long term trend in optical communications to double performance each year, reducing cost per bit transmitted.  The internet traffic and bits shipped are growing at steady exponential rates, so that future bandwidth, capacity and price performance can be anticipated.  Companies further down the food chain find themselves squeezed, and are looking at other markets that they can take their technologies to, with different products.  Optical Coherence Tomography and Endoscopy are two areas that is benefiting hugely, with spectral engines developed for monitoring communications becoming available as light sources.  Bayspec and Axsun are two subsystem companies moving into this space with a range of products.

Electronic 100GbE and Its Transport over OTN: Current Status and Future Developments

Confirmed what we’ve been seeing for a little while now, that 100G will mark a paradigm shift in the industry, leaving behind two level (1 or 0) transmission, and moving to multi-level formats using phase, polarization and amplitude properties.  Verizon talked about three field trials that they’ve done, using Alcatel-Lucent, Nokia-Siemens, and Nortel kit.  The word is that the Nortel kit handles polarization mode dispersion (PMD) really well, and in fact better than many 10G system.  Well done to the Nortel guys for their vision in this space – they’ve really set a trend for the industry to follow.  There was also some discussion that the other paradigm is also broken:  for previous bit rate increases the cost point has been a 2.5x increase in cost for a 4x increase in speed.  Now I’m hearing arguments that it’s only got to be cost-competitive or slight advantageous for many carriers to want to adopt. 

AT&T talked about implementation in core networks first (less cost sensitive) followed by metro.  They see 400G as the eventual ‘sweet spot’ from fundamental limits, so will be interesting to see if this eventuates.  They also talked about differences between carriers – the age of their fiber and it’s properties mean that the 100G equation is different for different companies.  For some, particularly those with older fiber that may have lower quality, going straight from 10G to 100G makes sense, while for others with good fiber, 40G is a natural intermediate step.  On evolution paths, they thought that bit rate-mixed systems make sense with 10/40G available now, and 40/100G coherent mixes likely in the future.

Beyond 10-Gb/s Passive Optical Networks – What’s Next?

There was a lot of discussion here about innovation in a strongly cost-competitive space.  There were some pessimistic figures put up about how much is spent in R&D in PON – sparked some strong opposing views from the floor and great debate.  There were a couple of paths presented as options for going beyond PON:  leveraging CWDM using low-cost (but never cheap!) filters, or leveraging mobile comm technologies such as FDM.  Most don’t see the TDM aspect surviving beyond 10G bit rates.  There were some other options talked about as well, to do with overlaying PONs of different speeds at different wavelengths.  But the cost in doing this, and a sensible evolutionary path for existing PONs still seems to be an interesting question.

Beyond the Door Step: Can Fiber Also Invade the Home?

I only managed to hear the Verizon talk in this workshop – this wasn’t rocket-science.  Keys are low cost, and low inconvenience to tenants, so fast installs using bundled drop cables.  One issue that does crop up in buildings is using fiber that you can bend without having big losses – there was some nice stuff presented here on a specially designed and ruggedised large diameter fibers (2.9mm) with tight bends possible.  It’s all designed to be simple enough for technicians to install, and PhD’s definitely not required!  I still think there’s an interesting question here of fiber vs wireless – both can make sense.  But from a personal perspective (and maybe this is just a knee-jerk response to media hype), I’d rather not subject my body to a larger dose of constant RF transmission than I have to.

Are All-Optical Networks Manageable?

This was a really interesting glimpse into possible futures.  There has been talk for a long time about transparent networks with dynamic path control, and a few different viewpoints were presented here.  The first problem is whether you can predict what will happen in these networks from lab experiments, and whether you can model close enough to reality without an insurmountable cost barrier in the kit you need.  For example, one trick that has been around for a while is using a recirculating loop to emulate long distance transmission.  The problem is that you go through the same component each loop, rather than a samples from a manufactured population.  This means that filters automatically align, but imperfections get amplified.  There is an interesting question on whether manufacturing spec statistics could be available and trusted, to predict network performance.

On routing, selecting an optimized path through a network with many possibilities is a complex problem to solve (and yet our GPS units do a good-enough job most days), but there seems to be some progress being made here in Europe on Diconet.  Setting up shared protection paths – and being able to provision new protection paths on the fly if a primary path goes down, raised some interesting questions about how much spare capacity would be needed in the network.  Impairment-aware routing is a strong research topic at the moment.  There was a fascinating discussion on optical performance monitoring, impairment-aware routing, and the new complex modulation formats and electronic compensation – it’s unclear how these will interact.  Watch this space in the next few years.

If you were at the workshops, feel free to add your comments below.

• Anthill magazine get involved in all sorts of general and targeted events, some free, some paid. http://anthillonline.com/category/events/
• The Hive is a networking club for entrepreneur, and run events, usually free,  in Melbourne, Sydney and Brisbane.  They offer free membership, and have podcasts of some of their events  http://thehive.org.au/category/melbourne/ 
• The Churchill Club describe themselves as ‘A community of technologists, entrepreneurs & innovators, who share ideas, wisdom and experiences’  They also regular events, and charge for events, which get more expensive the closer to the time that you book.  They have member and non-member rates, and offer a free weekly newsletter. http://www.churchillclub.org.au/index.php
• RMIT’s Enterprise group run a number of events, many of which are early morning.  Both their ‘Coffee Club’ and ‘Business Breakfasts’ are free, and there is a mailing list you can join to get advised of upcoming events http://www.rmit.edu.au/bus/enterprisegroup
• NICTA run occasional  ’Meet the Founder’ events, usually in the evening.  They are free, there is a mailing list, and you can get recordings of some past talks. http://www.nicta.com.au/nicta_events/meet_the_founder/vrl_series
• Design Victoria run a series of educational events, panels, and workshops.  They’re generally aimed at SME’s, but some of their evening ‘By Design’ events make good networking http://www.designvic.com/Events/DesignVictoria.aspx
• Startup Melbourne capture events from a bunch of sources, and offer it in a convenient calendar format.  If you’re visiting, this might be the best place to find out what’s on.  http://startupmelbourne.com/calendar/

I go to a bunch of these on a fairly regular basis – if you’re coming along, drop me a line and we can meet up.

 However, the Australian National Broadband Network (NBN) is well and truly underway, with a company created (unimaginatively, but descriptively, called NBNCo) that is currently going through an industry engagement process, and already making big decisions.  NBNCo has a single shareholder – the Australian Federal Government - and have a very strong sense already of who they are, what they will and won’t be doing, and what they want to achieve.  Mike Quigley, Gary McLaren, and Christy Boyce, the CEO, CTO, and Head of Industry Engagement, respectively, have been evangelizing the key features at various events.  Here are some of the goals they have presented, which make an interesting comparison against the Google initiative.

Coverage

The NBNCo mission is to get 90% of the population connected to FTTH, with speeds of (up to) 100 Mbps.  The remaining 10% will get connection speeds of  12 Mbps, by a mixture of wireless, satellite, and anything else that comes along that can get the job done cost-efficiently.  The brackets are significant.  The initial media release from Minister Conroy includes the ‘up to’ as shown here, but it seems to move around in people’s minds.  Nevertheless, Quigley has publically said that the 100 Mbps and 12 Mbps are their targets. 

In comparison, Google’s announcement is also for FTTH, but for over 1 Gbps, for between 50,000 and 500,000 people.  That could be single locations with many users, such as universities and schools; it could include high density, suburban, regional and/or remote areas.  Their stated goal is to ‘make Internet access better and faster for everyone.’  I think that the demographic they choose will speak volumes about who ‘everyone’ is.  They need to draw a careful line between trying to do all things for all people, and achieving nothing, versus testing a single demographic, that does not provide valid data to extrapolate into a business case.  Google have said that they want to test new ways to build fibre networks and help inform and support deployments elsewhere.  The Australian rollout is likely to be concurrent, so hopefully the testing and information-sharing may be two way.  Australia is likely to come across far greater geographic diversity – the country is the same physical size as the entire USA, but with a population of around 21 million, and densities largely determined by water availability.  If you’re interested, write a comment and I can tell you about some of the hideous things Australian soil can do to optical fiber…

Competition

NBNCo have said that they are building a Layer 2 wholesale only network (with a few Layer 3 features).  They are physically connecting the homes to FANs (Fiber Access Nodes) where there is existing competition, and then leasing the backbone capacity to create the central network infrastructure.  Each home connection will have 4 ethernet ports and an analog terminal adapter port, to allow for multiple providers. They have said that they have a very clear goal from the Government to occupy as small a footprint in the value chain as possible, and to provide a level playing field for competition, and that they feel this is the best way of meeting that goal.  Would-be retail (or indeed ISP wholesale) providers will still need to make an investment to get from Layer 2 to Layer 3, which creates some barrier to entry, and is thus a protectable investment.  It also allows scope for a lot of innovation in the sorts of services that can be provided.  The balance between barriers to entry, investment requirements, and the ability to innovate, is what led them to Layer 2 (and they had some very dense and funky charts to support this)

Google are well known for their support of ‘net neutrality’.  According to Silicon Republic, “Google have said that their network will also be ‘open access’, giving users the choice of multiple service providers.  Their intention is to see what next-generation apps developers and users can make use of with ultra high-speed and see whether it’s worth creating new bandwidth-intensive ‘killer apps’ and services or other uses that haven’t yet been imagined.”  Sounds pretty similar.  However, Google have not yet said how the apps developers will be engaged, and what level the platform will be that they’re offered (I know, give them a break.  These decisions take time and careful thought).  I hope like crazy that it will be similar to the NBN platform, as this offers a much more compelling case for developers to get on board, but time will tell.  And yes, someone did ask the question, and no, Google is NOT planning on doing its trials on Australian soil (link).  But you’d better believe that they’ll be watching closely to see what happens.

Summary

Google is a private company, yet with a public image of wanting to ‘do good’, and move beyond the corporate arena.  They live in the bleeding edge world of emerging technology, with the next unexpected killer application just over the horizon.  New business models abound in this space – theirs has been successful up until now, but they need to innovate to continue riding the wave.  Their attitude has been reported as, the more people that use the internet, the more money they can make.  People getting better broadband, and the benefits that brings, is almost a side benefit of this model.  The Australian Government wants to provide its population (including me) with nation-building infrastructure, to overcome the tyranny of distance, to improve the quality of life and opportunties, and to engage with (and take on) the world.  The government may make some money out of it indirectly, via taxes from a wealthier population, but they see the value of the internet in quality of  life.  Before you dismiss this, think back to life before every house had a phone, and how much that step changed the world.  So two different organisations, valuing opposite sides of the same coin.    I can’t wait to see what their new world brings.  If I get a better quality of life, and Google, or my Government, makes a little more money along the way, that’s OK by me.

For one way communications, such as television and radio, latency is not an issue – the delay is the same for all parts of the message, and you don’t generally notice that it is there.  For two way communications, such as video conferencing, gaming, or control applications, it is critical.

Brad Wearn from BHP Billiton talked about their use of broadband at Port Hedland in WA, which is a remote region of the country.  He talked about using broadband for remote control of equipment, amongst other things.  For these types of applications – which the NBN will enable, and will have huge benefits for industry – latency is critical.  BHP cared enough to lay a new fiber optic cable, as satellite latency was too high.

Fundamental Differences between Satellite and Fiber

Latency in satellite communications is fundamental – there is nothing you can do about it, as the signal has to travel all the way up to the satellite, and then all the way back to a receiver.  There is a really nice discussion about orbit heights on the Becta website here.  For Australian applications, the idea is to reduce cost for areas with low population densities, and so the satellite orbits will be higher – Geostationary orbits make the most sense to provide permanent coverage.  These are at an altitude of around 36,000 km.  So the round trip distance for a signal is around 72,000 km.  At the speed of light (3×10^8 m/s), this will take 240ms – for your signal to reach the destination.  For a signal to return to you from the destination is another 240ms.  This is with no allowance for network gateways – for my office in Melbourne this is 54ms, or for signal processing at either end to work out what to do with the data that has been sent.   So with satellite, you’ve got an irreducible half second delay (unless you’re planning to invent faster than light communications), which is irritating for video, and can make control applications unusable and unsafe (things happen faster than you can remotely react to them).

This is one of the reasons why optical fiber is a really attractive alternative, everywhere you have sufficient population density to roll it out.  Light travels down optical fibers slightly slower than the speed of light – at around 2×10^8 m/s.  Let’s calculate the comparative latency going back to the Port Hedland application.  Port Hedland is 1,640km from Perth, and so the travel time via fiber is 16 ms – return.  This is about 100 times less delay than the satellite link.  In practice, processing actually dominates the latency in these connections, and the improvement isn’t quite this good. 

Here’s another example.  Wikipedia says there is a Google Server farm in Mountain View, California, and another in Tokyo.  They’re both about 7,500km from Melbourne.  Round trip time via fiber will be 75 ms + processing.  This is still much, much less than satellite latency. 

There is a very nice website called speedtest that lets you see the latency from your location to other servers anywhere in the world.  From the figures, you’ll be able to work out that these connections are all optical fiber based.

Implications

What this means for remote areas of Australia with satellite communications, is that they’re never going to really be able to Skype properly, or take part properly in Webinars with interactive components.  For Australian Industry like BHP with complex and expensive remote operations, there will remain a strong business case for providing optical fiber links.  Hopefully, they will follow BHP’s lead in providing the low latency, high bandwidth connection to the local communities along the way. 

BHP have said that they’ve already found benefits in providing this service – in terms of making it easier to attract and keep good people at these locations.  I’ll be very interested to see further case studies on these implementations and their local community impact, as they happen

The focus of the forum was largely on the health, education and business applications that it will enable, and speakers had many stories and examples of the successes of understanding, measuring, and strategizing digital media.  For example, Aussie home loans told a fascinating story of their engagement with digital over the last 10 years, and how it has changed their business.

 Mike Quigley (NBN’s CEO, in case you’ve been living under a rock somewhere) did do one semi-technical presentation, which answered two of the biggest questions around the NBN:

• Why fibre, not wireless: Essential point here is that fibre incurs the cost once in infrastructure investment, and can grow to meet consumer demand for bandwidth for a long time to come.  Wireless would require further investment and more investment as bandwidth demands grow, and would end up being more expensive in the long run.  It makes a lot of sense to me that the Government is making a larger but controllable investment in infrastructure now that can’t be derailed in the future
• Why 100 Mbps: His point was that traffic demand has grown by a factor of about 10 times every 10 years, and there is no reason to expect this trend is going to change.  Anything less than 100Mbps is nuts if that’s the growth we expect.  He also pointed out that we’re one of five regional initiatives for bit rates of 100 Mbps to 1 Gbps – so we’re still not ahead of the region in speed, but we ARE ahead in getting to the population

One thing has become very clear through the forum – that most Australian business, and particularly larger Australian business, really doesn’t get it.  You have only to compare the websites for Amazon (www.amazon.com) with Dymocks (www.dymocks.com.au) or Macy’s (www.macys.com) with Myer (www.myer.com.au), or Target US (www.target.com) with Target Australia (www.target.com.au).  The differences are striking, in terms of user experience.  The US sites understand that you’re there to buy, and make choice easy.  The Australian sites are largely pointers to catalogue pages.  We already buy a lot of our books from Amazon, as the prices are cheaper, and the services is better.  I wonder how long it will be before far more of our purchasing goes the same way, and what our retailers are going to do to change the trend.

 One interesting reflection from the conference is that boards have a role in mitigating risk, as there seems to be a perception that doing nothing is a low risk option.  It’s not.  There is a clear case for CEOs and boards for having a much better understanding of what broadband means for their business.  Not the nuts and bolts, but for new business models, marketing stategies, competition, customer understanding and engagement, sales, and profits.

 Photos are at www.techbus.com.au/photogallery

Here are some things to think about when recruiting for a startup.  I hope that you find them useful.

1. Do we like them?  Not do we like their work, or do we like what they could do for us, but do we like them as a person? You have a small team in a pressure cooker.  Research shows that adding or losing one person in that group changes the whole dynamic.  You need to like the people you work with – after all, you’re going to be spending a LOT of time in their company. 
2. Can they do the job?  This is where people can make the mistake of putting all of their criteria.  It’s important, but nowhere near sufficient.  My approach here is to take someone technical people into the interview, and they’ll soon put them through their paces.  But often, the most competant technical person is not going to be the best value to you.  So set a minimum standard, and look at the other criteria for anyone who crosses it. 
3. Will they do the job?  This is different from #2.  It’s all about attitude.  Think about the fact that you probably don’t know yet the full skill set that you’ll need, and all your people will be learning new things to cover the necessary bases.  I look for signs that people are life long learners, and will be able and interested in climbing new learning curves. Think about the fact that you’ll be up against seemingly impossible deadlines.  Ask them what they’d do faced with a deadline they don’t think they can meet. 
4. Will they fit in?  Not considering this will make your life miserable.  Think about the culture you want to build, and what this person’s values come across as.  Trust your gut on this, even if you can’t verbalise your reasons.  And never, never, never fall into the trap of recruiting someone that is the opposite to you, because you think you’re weak in that area.  You’ll end up on opposite sides of a chasm, throwing rocks at each other.  Look for bridge builders instead.
5. Are we heading into groupthink?  If your company is all the same demographic in gender, age, and ethnicity, then you’re (a) missing out on the benefits of a range of perspectives in what you’re doing, and (b) likely to encounter some extreme behaviours.  It’s been shown that groups of a single gender are more likely to set up cliques, and exhibit exclusionary behaviour and have bullying problems.  Do yourself a favour, and set up a well-balanced workplace from the word go.
6. How much do they know about start ups?  If they don’t know what they’re in for, tell them up front.  Guy Kawasaki talks about this in The Art of the Start, and I’ve found it to be great advice.  Paint a rosy picture about how nice things will be, and they’ll feel betrayed when the s%^&t hits the fan, as it will.  Present a strong picture of how hairy things are likely to get and ask them if they’re up to the challenge, then they’ll feel that they’ve made a commitment to see it through, and respect you for having prepared them for it.  If they don’t take the job because if it, then they weren’t right anyway.

If you’ve found us directly through the blog site, then why not check out our website, at www.techbus.com.au