Like many others, I have been following the demand for Free Public Wifi on public transport with great interest. As Peplink’s Solution Architect I have been involved in a number of large scale projects recently where Public Wi-Fi in buses & taxis is not only in demand from the paying public, but needed by the vehicle operators to win new contracts, and from the perspective of complete solution design these projects have a number of really good technical and commercial challenges.

There used to be fewer options

If you had asked me how I would have provisioned customer Wi-Fi on trains 5 years ago the answer would have been easy – track side fibre connected Access Points with on-board high speed Wi-FI AP clients. The reason being that the only way to get enough bandwidth to a moving vehicle back then was via Wi-Fi, since cellular data was not only expensive but also much narrower bandwidth than today’s 4G/LTE services. In fact one of the first specialist Access Point products Peplink ever built was the Pepwave Express, a train  based Wi-FI router designed to do just this, with make-before-break dual Wi-Fi as WAN radios built in to cope with a high speed trains linear progress in and out of track side Wi-Fi Access Point Coverage.

Now we are all thinking about Cellular Data

Today of course things are different, with 4G/LTE services available nationally and improved hsdpa+ services in between, there is a lot more cellular coverage generally – its still not complete coverage though, and as this article from back in October of last year noted, Mobile coverage on trains really is pants. Certainly from my personal experience of journeys I have taken recently (both road and rail) frequently (and often for extended periods) the only cellular data connectivity available was GPRS and EDGE with the occasional not spot thrown in. LTE connectivity was certainly the exception and not the norm, so cellular driven public WiFi will need heavy infrastructure investment.

Where will the investment come from?

Last week David Cameron announced a commitment for free WiFI on UK trains by 2017, with a £50m pot of money released from the Department Of Transport for operators to ensure WiFI is available on a number of UK train routes, and so the hope is that with this new, heavily publicised investment injection, mobile service operators will look to develop and improve their own cellular networks and coverage in tandem alongside our nations rail and road networks.

What does the end solution look like?

With the hope that high bandwidth cellular coverage will continue to improve then – at least to the point where there is ubiquitous coverage across the national rail infrastructure, in vehicle cellular Wi-FI routers would appear to be the best mid to long term investment for train operators to consider.

In road based vehicles, like cars, taxis and buses, we have many customers choosing to install either singular LTE routers (like our BR1) or dual and quad cellular routers (like our HD2 and HD4) depending on passenger volume and specific applications. However with train based deployments a number of additional challenges and questions arise.

  • High End User Density, with many train coaches seating over 70 passengers, a train with 8 coaches can have 400+ potential service users.
  • High Client Device Density – general expectations would be a single device (perhaps a smartphone) per passenger, but many could have multiple devices like tablets and tablets for longer journeys.
  • Bandwidth demand / consumption – if a service level were to target an allowance of 1-2Mbps per client device, some trains would require as much as 1Gbps of cellular throughput
  • End User experience – Coverage will naturally vary along the route, how can customer expectations of the service be set and communicated to improve the perception of the WiFI service as a whole?
  • Physical Device Constraints – Trains are steel tubes, the positioning and location of both internal WiFI antennas and external cellular antennas will be very important.
  • Compliance – the physical installation of electronic equipment on rolling stock is highly regulated with all equipment and installers requiring certifications, there is also the challenge of approvals for through body openings for external antenna arrays.

User and Client Density Leading to high bandwidth requirements

The number of possible clients on a full train and subsequently the amount of bandwidth required can be hard to gauge. Personal experience of public transport installs suggests an average of 15-20% passenger usage of Free WiFi services, but this usage is likely to increase as more passengers realise there is Free WiFI available (and begin to expect it) so we should plan for usage to rise over time.

If we stick with 20% and assume a full train can have 400 passengers then we are looking at 80 active users, and if we have a target SLA of 2Mbps per user, 160Mbps of bandwidth required. If the average single LTE connection can provide 10Mbps (sometimes a lot less or more depending on location) we are looking at (on paper at least) 16 cellular modems to provide the raw bandwidth.

At Peplink, our largest cellular router is the HD4 with four internal modems, so in this hypothetical design we would be looking at four HD4s split evenly across the 8 passenger coaches which seems to have a nice balance to it – one router per pair of coaches. These HD4’s have POE support on the LAN which could be used to run additional POE powered Access Points to extend coverage between passenger coaches.

When more bandwidth is required, each HD4 has additional support for two wired WANs per router, and so we could add a pair of BR1 LTE routers to each HD4 providing us with 6 usable LTE’s per pair of coaches or up to 240Mbps of bandwidth on average (so supporting 120 users @ 2Mbps / user which is enough for 30% passenger usage).

Managing User Experience and Expectations

Of course not all trains run at full capacity all of the time, and in off peak times a Free WiFI user might have a noticeably better user experience of the WiFI network that in peak rush hour periods since more bandwidth would be available. So the question arises as to how best to manage customer expectations and individual customer bandwidth allocation throughout their journey.

Typically in fixed installations I would approach this in a number of ways. Firstly bandwidth allocation and capping for each user connection, if a user always gets a 2Mbps connection – no matter how much bandwidth is actually available then they will always have a similar experience of the service – a bit draconian perhaps but of course premium services can be sold to allow paid subscribers to consume more than the typical 2Mbps. I would also use web caching whenever possible, since most users will generally visit the same core social media and news sites and these are easily cached to both reduce the amount of public internet bandwidth consumed and also speed up page load times through local delivery of regularly accessed content.

In fixed installations though there is the presumption of available connectivity – since WAN failure events on fixed line services are fairly rare. In mobile scenarios though, loss of connectivity is much more likely. Yes this can be mitigated, and the HD4 offers dual sim slots per modem so that multiple mobile provider sims can be used to better guarantee internet connectivity as the vehicle passes in and out of coverage areas, but there will still be occasions where physical and sometimes geographical restrictions with cause disconnects (picture valleys and tunnels for example).

Offline Content is key

So what can be done about those situations? I think there are two sensible approaches. Firstly completely offline pre-cached content, this differs from dynamically cached internet based content in that it is selected and advertised to passengers as an alternative content source. This could be TV shows or movies as part of an infotainment solution in partnership with advertisers or the production companies themselves for example, acting as a form of in vehicle media library. Or it could be web content from key media outlets (like the BBC or local radio / media sources) that has been actively packaged for the operator (by those outlets) for offline access.

Connectivity Visualisation

The second approach is to make passengers aware of the connectivity status and bandwidth availability to change their usage habits. Trains tend to have a limited number of toilets on-board, and their availability is advertised with a visual indicator showing if they are engaged or not. Potentially the same thing could be done for the internet connection. Picture a display screen on a bulkhead in the coach that shows how many clients are connected, how much bandwidth is available and the current connectivity status – all live and real-time data. If I can see there is limited bandwidth, I’m not going to try and stream BBC iPlayer in full HD, or if I do then I don’t expect it to stream perfectly, and if I lose connectivity as we pass through a tunnel a quick glance at the display board will show me that internet connectivity has been lost and when it is connected again.

Cellular Data Costs are Going to Hurt – aren’t they?

The sad fact of the matter is that yes they really are. In early trials on buses I have seen average monthly cellular data usage vary between 80-150Gb. Ok, there was no web content caching or traffic optimization in place in these trials and both of those will reduce data consumption dramatically, but the trials were also limited by their novelty and uniqueness – or in other words, passengers weren’t expecting to be able to use Wi-Fi so didn’t necessarily do so.

Monthly cellular data charges for bandwidth at this scale will be painful for any operator to absorb for a service they are literally giving away. Government grants will help of course – at least initially, as would the capability to offer premium subscriptions to the small number of regular service users who can justify the outlay so they can work (or relax) effectively using bandwidth that has been reserved for their use. However for Free WiFi to work for operators at this scale, new revenue streams will be required to compenstate them for their monthly data charges – so where will this revenue come from?

Monetization of Free WiFi services

I see the Free WiFi services as just another delivery platform for media content, and believe that it will be the content producers (and their marketing budgets) that will likely be the genesis of the required revenue for operators to provide Free WiFI services. Media companies like, Sky, ITV, Amazon and Netflix who would be happy to offer a selection their back catalogue of content and taster episodes of new shows in return for the opportunity to demonstrate their services and quality of their content to a new potential audience. I would expect it to be a considerable opportunity for them to win new subscribers by offering discount codes and free trials to the train based audience, but ultimately only time will tell if that really is the case.

Naturally advertising generally has a place as well, with ‘airtime’ sold on captive service login portals and regular re-auth screens to the highest bidders, and custom operator information and WiFi ‘helper apps’ able to push even more marketing content too. In fact the combination of smartphone based apps and onboard WiFi will likely hit the sweetest spot, with user experiences tied together by the app and integrated into other services like geography based focused marketing advertising the latest available offers and deals at shops, restaurants and businesses as they fly pass the window.

What Is Peplink’s Approach to Public WiFi?

Our focus is (as it always has been) on providing the best possible connectivity everywhere its needed. We specialise in managing multiple cellular and over-the-air connections (like WiFi as WAN and mobile Satellite) on single devices, and are rightly proud of everything we have achieved so far in the MAX HD and BR range of products. These are used globally now in just about every vertical to provide ‘unbreakable’ VPN connectivity using our SpeedFusion technology and have proven themselves in the harshest of environments. 2015 will see the arrival of a number of really exciting new products designed specifically for Public Transport connectivity, enhancing operator and reseller capabilities in this space . However, as I have pointed out, awesome management of connectivity alone is just one piece of the free public WiFi puzzle.

Technology Partnerships

With the maturity and coverage of LTE services in the UK potentially reaching a tipping point this year where availability of LTE services along the full length of rail and road journeys will likely become a reality; one area of focus is on building the relationships required with media companies, advertisers, and monetization services that complete the solution design circle and actually enable the successful commercial delivery of Free Public WiFi at repeatable scale.

I believe that ubiquitous LTE cellular coverage combined with the ability of Peplink technologies to intelligently and elegantly use it, in partnership with media, marketing and advertising companies has the potential to deliver a truly ground breaking solution for Free Public Wifi everywhere, and I am looking forward to seeing what will happen in the coming months.

One thing is for sure, not matter what happens, its going to be really exciting year for Free Public WiFI.