Hidden costs

Using a private mobile network brings a lot of advantages. Most of this website is dedicated to that in one way or another but primarily they are;

  • security- keeps the data and the device off open internet
  • lower costs- fixed IP means less resends and avoids the costs of being spammed and spoofed
  • visibility- a decent private network should offer you better diagnostics and alerts that you can usually get off a Mobile Network Operator (MNO).
  • But most Mobile Virtual Network Operators (MVNO) don’t charge for their backhaul. This was fine when you had 100,000 SIMs sending one 9byte meter reading each a week. The model, and the private network, breaks when you start to stream 50 HD cameras through a private network in real time on 4G connections and you have to buy another 1gigbit per second capacity just to support those hungry few.

    But it is more than the considerable cost of the pipe. If like Mobius, you run a full dual location service then you need two pipes so that the data throughput isn’t impacted if one fails. The servers and firewalls in both locations need to be upgraded and there seems to be an endless number of licences that need expanding to cope with the ever-increasing volume of transactions, connections and data.

    But if you are serious about offering a Service Level Agreement (SLA) then, for example, what happens if one of the servers’ crash? You are back down to one connection. Your managed service provider will swap out the hardware on a four-hour break/fix, another cost, but how long will it take you to rebuild the servers – certainly hours, probably days.

    Mobius added extra servers at each point in both primary and secondary links so that if a server crashes, then a mirror is immediately available for the rebuild, on location, where it needs to be. We then built a complete duplicate system of all the servers, Cisco routers, firewalls and connections. Last week we moved our IT Team offsite for two days so that they could war-game with a third party company working through a break in each system, not just the data pipes but the back-ups, the hardware, the firmware, the software that we use to run our company, our billing platform and our diagnostics.

    The results were better than we hoped and highlighted a couple of the older internal systems that we use where probably some work is needed or be replaced.

    But next time you entrust your data to a company that gives you the backhaul away for free ask them for a copy of their Disaster Recovery (DR) plan.

    A SIM is for life not just for Christmas

    A SIM is for life, not just for Christmas. We have all got used to the networks sending out SIMs as part of an upgrade or swapping network. The user takes the back of the phone off, pops the new SIM in and away you go. Even the introduction of a dangerous looking crimping tool to get the pico-shaped SIM your new phone needs is taken in your stride. You can be confident that your cherished phone number will follow you from network to network under clearly defined and understood Ofcom rules.

    In this world of transferable business picking an airtime provider is a relatively low risk process. If it turns out the technical support is poor, the SIMs are consumer grade or the backhaul is so contested you are losing data then not all is lost. Simply transfer the SIMs to a new provider and everything will be better.

    But M2M is different. By definition the SIM is in an unattended and possibly hostile location. It may be potted into the unit, making a physical swap impossible. A Port Authorisation Code (PAC) does still allow you to transfer the number but in these days of IP that has little value. The phone number, if it exists at all, has no commercial relevance. The M2M SIM, once it is in a unit, is almost certainly there for the life of the product, even before you start to look at the contract.

    So, when you are reaching for that packet of SIMs, or buying the cheapest airtime you can find on the internet, pause for a moment. How long is that unit going to be out in the field? What is your planned mean time between failures, recovery cost, and loss of revenue for failing deliver the service? If the M2M provider isn’t charging you for industrial grade SIMs, backhaul, technical support and even customer service what kind of shape are they going to be in when you need them in two three- or ten-years’ time?

    When thinking about buying a SIM then demand to see the Network Infrastructure, ask to see the DR plan, the Service Level Agreement, meet the customer service team and have a look at the billing information. Make sure you get SIMs that are for life, not just for Christmas.

    Donald Rumsfeld was right

    “…because as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns — the ones we don’t know we don’t know.”

    The former Secretary of Defence may not seem the obvious person to have an insight on the issues facing Machine to Machine (M2M) or the Internet of Things (IoT) at the moment. However, one of the issues that we see again and again in mobile is the following pattern-

    The system – ticket machine, car tracking system or traffic light control is built, usually as a custom build with a group of disparate parties who may never have previously worked together. Everyone should start with a set of known knowns- that is they know that their bit of kit works (but this is nothing as like as common as it should be.)

    The system is labbed up. The system does not work. All parties involved check and recheck their component(s) in the system to make sure they have addressed all the unknowns that they know about.

    The System starts to work but since so many parameters have changed no one can be sure why it didn’t work but now it does. Everybody freezes but at that point there are unknown unknowns. Because no one knows what conflict stopped it working and then went away, no one can know what change of circumstance will stop it again.

    But it works. And there is a deadline, and the product is then rolled out. And then it stops, sometimes. At that point the circle of blame begins because each part is known but the system is so full of unknowns that the behaviour seems almost organic in its responses.

    There are at least 8 factors affecting how the SIM alone relates to a network. Unless all of those are right and tested at point of shipment, and most aren’t, then it won’t work, or won’t work some of the time, or gets kicked off, or the data going through it gets choked when it gets busy. And then you have the Mobile engine, the firmware, the AT commands, the mast, cell breathing, the filters, the self-protection, the GGSNs, loading, VPN and you haven’t even hit your firewall yet.

    Mobius always starts from a known position. We know our SIMs are correct now, in six months’ time and in six years. We know that our private networks are working in real time. That concrete certainty means that when something isn’t working, we will know what it is. A known unknown either Downstream or Upstream but with Mobius as that known good state in the middle.

    If you are rolling out an M2M or IoT system, then make sure you are dealing with a company that knows what they are doing. 

    Diversity – a forgotten radio concept

    Diversity – a forgotten radio concept. So much of Radio Frequency (RF) performance was a dark art when there were still lots of people that understood it. It is not then surprising that 20 years into a world of radio modules offering IEEE standards with an easy ‘AT Command’ set of instructions that the level of understanding on RF may have reached a new low. Many have forgotten that Bluetooth, Wi-Fi and even cellular itself are even radio systems, especially since they took antenna off mobile phones.

    It is easy to overlook that the mobile signal that your phone is receiving may actually be taking a complex path, bouncing off passing buses, the foil insulation in the wall, the glass panel in your meeting room before reaching your phone. Usually all of the above at the same time. As all of these tangles of signals phase in and out it creates dead spots where only a step away reception is loud and clear. Subconsciously we know this and often wander around a room while on the phone hunting for that magic spot where you really do get good reception. For a few moments at least.

    But in Machine to Machine (M2M) we don’t have the luxury of moving the unit around in real time to optimise its position. In M2M that’s why antenna performance becomes critical in the overall performance of the unit- the M2M kit is very rarely placed with RF in mind. It is where it needs to be because of all the other factors- power, security or just ‘the only space left’. The antenna needs to make the most out whatever radio reception is available.

    It’s why Mobius won’t sell airtime to somebody using USB Dongles. The antenna is such low gain and usually fairly limited directional performance (as well as the consumer nature of the electronics inside) that there are simply too many issues to overcome to make the system reliable.

    So having an external antenna is a really important start. When Mobius is then asked why this is a better terminal rather than something cheaper, part of the answer is often that it supports a Diversity Antenna. So not only does it have a good external antenna, but it gives you the option of adding another second antenna a critical distance apart, catching the radio wave at a different point in the cycle. It allows the terminal to unscramble all of those reflections and build one ‘good’ wave from the different signals at the two points. In short, a Diversity antenna may work when a single one can’t.

    So up until the point M2M can get up and walk around the room looking for a better signal, it is worth looking at the hardware and asking how can I make the most of what is there at any given moment. Using Diversity Antennas, like using Mobius airtime, is another piece of the puzzle in building a more reliable M2M system. 

    Is voice roaming bad for business?

    Is voice roaming bad for business? SIMs that can roam in country are a key part of Machine to Machine (M2M). For example, an O2 SIM that can also use EE and Vodafone if there happens to be no O2 available. This is very attractive- it can create a more robust mobile connection for that last over air 2 or 3km. It cuts down on survey costs and can make preparation for a large scale roll out much easier. Once the estate is out in the street then swap out becomes easier and billing and backhaul should be unified- lowering running costs.

    One of those few occasions where you get better for less. So why not apply that to Voice?

    The pull of being on network wherever you are is a big one. Despite all the arguments the Networks threw up against the Governments plans, they largely missed out the most important one; If access is available to all then there is no incentive to invest in the network. Everyone would simply buy the cheapest airtime they could find because once that network is swamped due to lack of investment, they could simply roam on to a network that wasn’t swamped. Government forced roaming would mean that there was no return on investment in network performance or capacity. It all would quickly be reduced to the lowest common denominator- a nationalised mobile network where price is the only factor and dissatisfaction is high.

    Why then is in country roaming allowed in M2M?

    First of all, it there are controls that are put in place to regulate roaming. At its crudest some roaming contracts are simply overseas SIMs brought into the country that can roam on all networks the same way your phone can when you go abroad. The danger here is that the original network may have a rule that unless the SIM returns ‘home’ after say three months then the SIM is disconnected. Even if it doesn’t, there has been a rise in the last year of the ‘roamed onto’ network recognising that a SIM has outstayed its welcome and refusing access after a period.

    Even with full M2M Roaming SIMs there are still rules such as PLMNs for pointing connections onto preferred networks by the SIM. These can be got around by hardware configuration. But you also may experience “Steering” applied by the home network which will have an impact as it is outside of the control of the SIM, the hardware and the roaming network.

    The other restriction is that the pricing structure supports low data usage – 256k to 2mb a month. Sometimes this is stretched to 100mb or in some cases 500mb per SIM per month. Gig roaming or bigger bundles are normally the province of Mobile Broadband tariffs with all the usual restrictions, chokes and filters that a fair usage policy brings.

    Put all of these together and the controls are already in place so that no M2M SIM becomes too much of a burden on the host network. Making the most of those controls, making sure you don’t fall foul of them is something that Mobius has more experience of than anyone else in the UK- we supported a roll out of a full large scale Urban Traffic Management and Control system with Real Time Passenger Information all on roaming SIMs back in 2011. It meant that Dorset CC could cope with the 300,000 extra people coming in every day for the Olympics utilising any available capacity on all of the networks. The project never made the headlines because it all worked.

    Voice would have a very different experience. 

    And my wife’s recipe for ratatouille

    We have all got used to Mobile Contracts. There may be pages of script so small that an arc lamp and industrial strength glasses will not be enough to allow the thing to be read. Online where you can make the font any size you want, we still skip past the text to the “Accept” button. Some of the sites don’t even have the pretence you have read to the bottom and cheerfully offer the button right up at the top, highlighted and ready to go.

    How can it be that we all become so blasé about committing to very serious legal requirements with what are very large and powerful corporate entities? The clue comes in part to that very disparity of the size of a consumer versus the multinational corporate. The negotiation is inherently unfair, the company has the money, the lawyers, has spent months, if not years drafting the document that is put in front of you to sign with minutes or seconds to take in and commit to.

    We can sign that contract because, in broad terms, there is an assumption that the corporation will have a duty to be ‘reasonable’ in its dealings with its customers.  It is not reasonable to ask for your first born or, more likely, to ask for £52,000 worth of overage. The consumer is unlikely to understand the full or potential impact of their actions, even if it is outlined in Paragraph 6.1.1.2 of the contract that they have signed. Such clauses then become unenforceable and may even jeopardise the validity of the whole contract.

    The situation is very different in a company-to-company environment. Here Ofcom, who are aggressive in defending consumer rights appear to have an attitude that a company should be big and ugly enough to understand the commitments a company has made in a contract. It is vital then then that an airtime contract is read and understood before signing. What are the real costs at the end of the contract? How can that contract end? How are disputes settled? What is the role of the networks in an M2M environment?

    So, when presented with a Company to Company contract for airtime it is vital that we apply more caution than the usual cursory glance. It may well contain unpalatable things, and it may be that these items are there for good reasons. But if it does include a clause that you have to hand over your recipe for ratatouille beware! It is unlikely that Ofcom will be interested if later you decide that is unreasonable. 

    *Subject to fair usage policy

    *Subject to fair usage policy?

    Is Fair Usage fair? We discussed previously the problem that networks have with consumers and the impact that has on enforcing contracts. The imbalance in the relationship between the mighty networks and the ordinary Joe on the street. This imbalance means that the consumers have a blank cheque from the network- they can go to any country, stream video 24 hours a day or sign up to those dodgy text services. When the network attempts to charge for the services already provided, they are told that it is “unreasonable” to expect a consumer to pay tens of thousands of Pounds to cover the costs incurred.

    The networks needed a defence mechanism. If they can’t (broadly speaking) charge, what can they do? The answer appears to be the Fair Usage Policy. This old catch all phrase has been used to cover a very broad spectrum of activity.

    The most obvious one was the asterisk at the end of the statement “Data bundle- 5gigs*”

    *Subject to fair usage policy.

    What that really meant is that your data is 5Gig but what is likely to happen is that when it looks like you are going to exceed that data allowance then the network reserves the right to choke the throughput of that particular connection.  You’re not getting cut off, you are just getting slower to make sure that you don’t exceed the bundle and start racking up serious overage. It can even mean a hard stop if the throttling doesn’t seem to be working.

    But Fair Usage can mean much more. It covers the networks responsibility as a deliverer of content. DRM, Porn, even the skin tone content of pictures can be flagged. If the file is a picture, then the network reserves the right to compress it in order to keep your usage down and save you money.

    It can also cover how the SIM is expected to be used. In the networks modelling – a Mobile Broadband (MBB) SIM should be used in a USB Dongle for an hour or so in a café to download emails and catch up on the ubiquitous Facebook. This sensibly extends into a profile of non-continuous use, no peer to peer, no extended VPN use and so on.

    So, the networks can now offer a large seeming data bundle safe in the knowledge that they know how and when it’s going to be used and the comfort of knowing that even if they can’t bill for overage and miss-use they have a back stop, they have a way of managing the risk. This also allows them to hit pricing expectations.

    The consumer at the same time can be reassured that although they have a responsibility to keep the SIM safe and to manage their own usage the best, they can they are not risking bankruptcy every time they pull out their phone. Both parties win. Fair usage does indeed appear to be fair.

    Next time we will have a look at the unintended consequences of Fair Usage when it has been unwittingly applied to the world of Internet of Things.

    Can the Internet of Things be ‘fair’?

    Can the Internet of Things be ‘Fair’? We have looked at the impact of contracts and the rise of Fair Usage as a sensible response to the rather counter intuitive situation of a Network needing protection from their customers. This works to a degree of satisfaction in almost all the cases where airtime is used. Except one-

    In Internet of Things (IoT) applications (for the purposes of this argument I am going to regard IoT and Machine to Machine (M2M) as synonymous but for more details please have a look at https://youtu.be/QPl5Ts0Ygg4) the unwitting utilisation of Fair Usage SIMs has had a huge negative impact on mobile adoption.

    Remember Fair Usage means no peer to peer, no extended VPN’s, no streaming- The very things that you are likely to do in an IoT application. While you can establish a peer-to-peer connection and stream data, if you do it for an extended period the network starts to notice. The longer you run it or the busier the network the more you are likely to start to see disruption.

    Similarly, IoT Apps are built with a requirement for throughput and latency. But remember as the month goes on Fair Usage means that you are likely to see chokes, throttles or traffic shaping. If you are experiencing this kind of behaviour this may point to backhaul issues, but it could simply be that the network is applying its Fair Usage policy.

    The last problem is the most subtle of all. Fair Usage usually includes a line about “Occasional Use”. Remember that the network sold the SIM with the picture of someone drinking their Double Skinny tall caramel latte with Marshmallows while they check their emails. So, if a SIM has been on for three weeks continuously streaming NOX levels it is in breach of the policy it was sold under.

    The network is then acting quite reasonably if it kicks the offending unit off network. The unit then knows it should be on network and attempts to reattach. The network sees this unit trying to reattach despite it being told to go away and interprets the reattachment as an attack. Things go downhill from there.

    But how would you know? The unit worked fine in the lab but now it struggles to stay on network, gets constantly kicked off and even when it does get on fails to get anything like the connection speeds you were expecting. You might blame the network, the hardware vendor or even your own developers.

    You can specify Machine to Machine SIMs, but many vendors confuse Mobile Broadband (MBB) Business to Business (B2B) with M2M as all being ‘data’. Mobius has only ever sold M2M Airtime so provenance with us is beyond doubt. If you don’t, the next time it stops working it could simply be that the network is trying to be ‘Fair’.

    The end of mobile?

    There is a raft of new radio technologies coming out. There usually is but this time they are coalescing around Internet of Things and it appears to be adding a bit of energy and excitement into the sector. Internet of Things isn’t new. Here we have been selling the idea since 1998. Things really took off in 2001 with wide scale adoption of GPRS and 56kbp/s became a reality!

    One of the things that struck me with this new found focus on IoT is that there are a host of technologies either coming to the fore for the first time, or being dusted off because now they will have their time. At last people will recognise the advantages that this or that radio standard brings.

    Whatever the radio system- PMR, 4G or proprietary short-range radio there are some simple factors that come into play- How much data needs to be transmitted, how far does it need to go and how long do the batteries need to last?

    Cost is a function of these parameters plus volume- the more people who buy into this, the quicker the costs are driven down. Does it need a supporting mast network? If so, who maintains them, does it need a licence, do you need national coverage or will it be site by site and so on?

    The idea of building a private network where you are in control, where you don’t need to pay someone else for every data byte transmitted and you don’t have to worry about what other people are doing that could bring the whole network down are very attractive. When you couple that with the frustration generally felt with existing Mobile Networks regarding billing, downtime, latency and support. It starts to feel like anything would be better.

    But building a national network is expensive. Siting masts can be difficult, ensuring 24-hour access and maintenance, support, fault finding, power, security upgrades and expansion has a real cost. By the time you have implemented this you end up looking like a Mobile Network Operator. Even a strong technology like MESH is dependent on the quality and density of the network on implementation and then on support and upkeep afterwards. Quality costs.

    However, as we have seen in the right hands (Mobius), Mobile can deliver 99.996% availability on street in industrial applications. It can provide a very cost-effective backbone to any IoT system with international reach, low hardware costs and broadly understood principles. Does that mean that Mobile is right for every application? The answer is probably not. The rise of the Mobile Platforms from Apple and Android have changed the world of data gathering and information delivery. Adoption of architectures such as Swarm, Mesh and Fog have stretched the art what is possible well beyond putting a SIM in every device.

    It seems likely then that local data collection will be built using a broad variety of radio technologies depending on power consumption, distance and throughput.  When it comes to collecting that data from anywhere in the world and delivering it to your cloud try Mobius, it’s built, it costs less than you might think, and it delivers.

    How secure is mobile?

    There are various scare stories about Security on the internet in general and mobile in particular. This seems to provoke three reactions-

  • Ignore it as hyperbole, the internet is so big no one will find me
  • The data itself doesn’t contain anything useful so I don’t have to worry
  • I will never put any data on mobile or internet as none of it can be trusted
  • I am always struck by companies who have very strict rules on how their IT systems connect with the internet in the fixed world. When it comes to Mobile however, they apply some or all of these rules and in the process throw out everything they have learned about security. Let’s have a quick look at those points again.

    Ignore it as hyperbole, the internet is so big no one will find me

    This first one is easily addressed. If you have a look at Shodan here is a product developed specifically to find M2M devices, unprotected, out there on the internet. Twenty minutes is all you need to start finding devices with surprisingly little protection as Mark Ward of the BBC found out in this article http://www.bbc.co.uk/news/technology-22524274 . Hiding in the herd is no longer an option.

    The data itself doesn’t contain anything useful so I don’t have to worry

    This is a statement that we come across repeatedly. That the data itself has no value so why would somebody bother? This misses a couple points:

    the rise of bots- these are automated programmes usually on unknowing hosts that hammer away at systems with poor security on the principle that something interesting may turn up. You will be hacked to see if you’re worth hacking rather than hacked because you have something of value they are after.

    On mobile you pay for all the data going up and down. So even if your device is secure, or the data that it is carrying is eventually viewed as having no commercial value you will be in the relatively unique position of paying for your own attack.

    I will never put any data on mobile or internet as none of it can be trusted

    This simple solution means turning your back on the rise of Machine to Machine (M2M), of the Internet of Things (IoT) and perhaps the most important disruptor of all Smart Phones and their apps. It is also an overreaction. Data can be sent securely over the internet and mobile, often more cheaply, certainly more scalable, than the insecure methods chosen now. Over the next couple of weeks I will try and address how- at a SIM level, at a network level and also how you access that system.

    If you are interested in making Mobile more reliable, fit for industry, delivering the data you need please have a look around the other blogs or contact us at [email protected] or call us on +44 1530 511 180

    « Previous Page