The Internet of Things – trends for the telecoms, data centre, and utility industries

I gave the closing keynote at an event in Orlando last week on the topic of The Impact of the Internet of Things on Telcos, Data Centres, and Utilities.

The slides by themselves can be a little hard to grok, so I’ll go through them below. I should note at the outset that while many of my slide decks can be over 90, or even 100 slides, I kept this one to a more terse 66

And so, here is my explanation of the slides

1. Title slide
2. A little about me
3. The IoT section start
4. IoT has been around for a while, but the recent explosion in interest in it is down to the massive price drops for sensors, combined with near ubiquitous connectivity – we’re heading to a world where everything is smart and connected
5. According to the June 2016 Ericsson Mobility Report [PDF], the Internet of Things (IoT) is set to surpass mobile phones as the largest category of connected devices in 2018
6. Depending on who you believe, Cisco reckons we will have 50bn connected devices by 2020
7. While IDC puts the number at 212bn connected devices. Whatever the number is, it is going to mean many devices will be creating and transmitting data on the Internet
8. What kinds of things will be connected? Well, everything from wind turbines (this is an image from GE’s website – they have a suite of IoT apps which can “improve wind turbine efficiency up to 5%” which in a large wind farm is a big deal)
9. Rio Tinto has rolled out fully autonomous trucks at two of its mines in Australia. They developed the trucks in conjunction with Komatsu. The trucks, which are supervised from a control room 1,000km away in Perth, outperform manned trucks by 12%
10. A nod to one of my favourite comedy movies (“See the bears game last week? Great game”), while also introducing the next three slides…
11. Planes – according to Bill Ruh, GE’s CEO of Digital, GE’s jet engines produce 1TB of data per flight. With a typical plane flying 5-10 flights per day, that’s in the region of 10TB per plane per day, and there are 20,00 planes – that’s a lot of data. Plus, GE is currently analysing 50m variables from 10m sensors
12. Trains – New York Air Brakes has rolled out a sensor solution for trains, which it says is saving its customers $1bn per year
13. And automobiles – in the 18 months since Tesla starting collecting telemetry data from its customers’ cars, it has collected 780m miles of driving data. It is now collecting another 1 million miles every 10 hours. And the number of miles increases with each new Tesla sold
    And since 2009 Google has collected 1.5m miles of data. This may not sound like much in comparison, but given its data comes from Lidar radars, amongst other sensors, it is likely a far richer data set
14. With the rollout of smart meters, UK utility Centrica recently announced that it will be going from 75m meter reads a year, to 120bn meter reads per annum
15. Wearables, like the Fitbit now record our steps, our heartbeat, and even our sleep
16. This was my heartbeat last November when I presented at the SAP TechEd event in Barcelona – notice the peak at 2:30pm when I went onstage
17. Lots of in-home devices too, such as smoke alarms, thermostats, lightbulbs, and even security cameras and door locks are becoming smart
18. Even toy maker Atari has announced that it is getting into the Internet of Things business
19. Which is leading to an enormous data explosion
20. In 2012 analyst form IDC predicted that we will have created 40ZB of data by 2020
21. In 2015 it updated that prediction to 75ZB
22. Where will this data be created?
23. Well, according to the 2016 Ericsson Mobility Report, most of the IoT devices will be in Asia Pacific, Western Europe, and North America
24. When?
25. That depends, different devices have different data profiles for creation and consumption of data, depending on geography, time of day, and day of year
26. And why?
27. Because, as Mary Meeker pointed out in her 2016 State of The Internet report, global data growth has had a +50% CAGR since 2010, while data storage infrastructure costs have had a -20% CAGR in the same timeframe
28. In 2011 EU Commissioner Neelie Kroes famously said that Data is the new gold
29. And if that’s true, as is the case with any gold rush, the real money is to be made supplying the prospectors
30. Now, let’s look at some of the trends and impacts in the telecoms industry
31. From Ericsson’s 2016 Mobility Report we can see that the big growth for the telecoms is in data traffic
32. And not content to be merely infrastructure providers, telcos are looking to climb the value chain
33. To facilitate this data explosion, telecom companies are building fatter pipes with LTE growing significantly in numbers between 2015 and 2021, while 2019 will see 5G kicking off
34. Telcos are now offering cloud solutions. Their USP being that their cloud is fast, reliable, and end-to-end secure
35. There are huge opportunities for telcos in this space
36. In the next few slides I did a bit of a case study of AT&T, and some of the ways it is leveraging the Internet of Things. First off AT&T has partnered with solar company SunPower to connect residential solar panels for remote monitoring of the panels’ performance
37. In its connected vehicle portfolio, AT&T manage the connections for Tesla, Audi, GM, and Uber. They have 8m connected cars atm, and expect to grow that to 10m by the end of 2017
38. And, an interesting data point to back that up – in the first quarter of 2016, in the US, 32% of all new cellular connections were for cars. The largest percentage of any segment

39. 243,000 refrigerated shipping containers connected through AT&T

40. AT&T have a partnership with GE for intelligent lighting solutions for cities and public roadways
41. In the equipment and heavy machinery space, nearly half of all tractors and harvesters in the US are connected through AT&T
42. While in healthcare, AT&T predicts that wellness tracking and virtual care solutions will reach 60m homes & 74m users by 2019
43. Then there’s outdoor advertising. AT&T knows data analysis. For years they owned the largest telemarketing organisation in the US. Now, with cellular data, they can completely transform outdoor advertising. Previously for advertising hoardings, the amount of footfall, or vehicular traffic passing a sign could be guesstimated, but no more info than that was available. But now, because AT&T knows where everyone is, their gender, age, and approximate income, they can transform this business.
    Recently they carried out a study with a customer who wanted to advertise to women in the Dallas area who earned over $75,000 per year. They queried the data and found that the customer only needed to buy two billboards in all of Dallas, to adequately cover the target demographic. Needless to say the customer was impressed
44. Because they don’t have a monopoly on ideas, AT&T have opened up their M2X Internet of Things developer platform to allow outside developers create solutions using AT&T’s infrastructure
45. They’re far from being alone in this – Verizon have an Internet of Things platform as well called ThingSpace Develop
46. While t-mobile has announced that it is teaming up with Twilio for its Internet of Things play
47. And it is not just cellular technologies they are using – there are also other low bandwidth radio protocols such as Lora and Sigfox which the telcos are looking at to broaden their reach
48. I spoke to a senior exec at a telcom firm recently (who for obvious reasons preferred to remain unnamed) and he told me:
    “Telcos want to own everything, everywhere“The internet of things is certainly one way for them to get there
49. How is all this impacting the data centre industry?
50. Well, in the next four years data centre capacity will need to increase 750% according to IDC. Also required will be significant ramp-ups in analytics, security and privacy
51. As Jim Gray pointed out in his book The Fourth Paradigm:

    “As datasets grow ever larger, the most efficient way to perform most of these computations is clearly to move the analysis functions as close to the data as possible”

    In other words, instead of bringing all the data back to the data centre to be processed, more and more of the analysis will need to be performed at the edge

52. As a graduate biologist, this reminds me of the reflex arc – this arc allows reflex actions to occur relatively quickly by activating spinal motor neurons, without the delay of routing signals through the brain
53. So there will be a greater need for event stream processing outside the data centre – this will bring about faster responsiveness, and reduce storage requirements
54. This also explains the rise of companies such as EdgeConnex – companies who provide proximity, and lower latency
55. And the rise of new designs of racks for hyperscale computing, such as the 150kW Vapor.io Vapor Chamber which, according to a study conducted by Romonet is $3m cheaper per MW and reclaims 25% of floor space
56. Other initiatives in the industry include Google’s attempting to create a new standard for HDD’s to make them taller, adding more platters, and thus increasing IOPs
57. Microsoft and Facebook are getting together with Telefonica to build a 160TB transatlantic fibre cable (the largest to-date) to handle the vast streams of data they see coming
58. While Intel are warning that organisations need to become more security aware, as more devices become connected
59. I also decided to address a trend in data centres to require renewable energy from their utility providers, and did so by referencing this excellent letter from Microsoft General Counsel Brad Smith on the topic (recommended reading)
60. Finally, what about the utilities sector…
61. Well, there are many ways the internet of Things will impact the utilities vertical, but one of the least obvious, but most impactful ones will be the ability to move energy demand, to more closely match supply. If you’re curious about this, I’ve given 45 minute keynotes on this topic alone
62. Another way the Internet of Things will help utilities is renewables management (such as the GE example referenced earlier), and preventative maintenance applications
63. And finally, energy information services will be a big deal, for everything from remote monitoring for seniors, through to device maintenance, and home management
64. The conclusions
65. Thanks
66. Any questions?

I received extremely positive feedback on the talk from the attendees. If you have any comments/questions, feel free to leave them in the comments, email me (tom@tomraftery.com), or hit me up on Twitter, Facebook, or LinkedIn.