Of the many potentially interesting tracks being run at this year’s Cloud & Devops world, I’ve decided to focus on IoT and Data Analytics. This is an area of massive interest to RNF’s client base, as the apps we’re creating for them have the capability of generating novel and powerful insights. It’s an area that’s progressing at a rapid pace, so I look forward to hearing what the latest thinking is. Having said that, I might just pop over to the Enterprise Mobile track a little later…
Track IoT & Data Analytics
Chair’s Welcome, Emil Berthelsen, Machina Research
Emil opens by noting that new uses of captured data at source is changing things and that’t we’re living a highly disruptive period at the moment. Data & analytics is changing business processes. There are predicted to be 24.9 billion connected devices 2024. We will be talking of not peta- but yottabytes of data.
New game-changing element in the game: “fast data” which, in combination with “big data” is the thing to watch out for.
Talk 1: IoT Hype and Reality
Mal Minhas – Head of Software, Vertu Corp.
(Vertu makes high-end i.e. luxury handsets.)
On the subject of hype: Minhas cites PWC, Goldman papers, which refer to multi-trillion industry. However, emerging tech hype-cycle (Gartner) shows IoT at top of hype cycle; implication: it’ll be all down hill from here!
Minhas suggests we learn from predictions made in late 1990s about bluetooth. In 2016 we still haven’t attained the predictions made for 2002.
Describes commoditisation/innovation cycle:
Mal presents a definition of IoT:
HOWEVER: IoT is an evolution of proto-IoT technologies with 2 new elements: Sensors + Internet. With these two elements, what was likely hyped too early, will now come to pass.
Mal now describes his perception of the architecture, which makes use of sensors + the internet (sensors shown on the left):
…with some concrete technology examples…
Mal emphasises the “three Ss”: Security, Scalability, Speed
– Yes, it’s hyped
– But it will eventually land everywhere
– You can anticipate and prepare
— 5 element architecture map
— The 3 Ss
– Interoperability is one of the big challenges of IoT. Hypercat is one attempt to address this.
– Difference between Consumer IoT and Enterprise IoT. Mal gives one example, which resulted from a survey: Vertu customers (consumers), for example, are concerned about security of IoT.
Good opener. Necessarily high-level, but definitely sets the scene for things that I want to hear about today and tomorrow. If I have one criticism of Mal’s content it’s that his approach appears to be very technology driven as opposed to being driven by business need. It’ll be interesting to hear what kind of emphasis the other speakers bring as the day proceeds.
Talk 2: IoT will Make Digital Enterprise *More Secure* [Mike’s emphasis]
The IoT Security Paradox
John Miri, LCRA
Miri is involved with power generation in Austin Texas:
LCRA was a leader in bringing tech (electricity) to rural Texas. IoT is apparently the new electricity, which brings…
Exponential risks vs exponential opportunities: “We need to make sure the opportunities outweigh the risks.” [Mike: Yes, but what if the risk is a nuclear bomb e.g. like personal privacy becoming a thing of the past?]
John claims risks a driven to a large extent by the growth in IoT devices:
Additional risk: There are competing standards. In other words, there is no unifying standard for IoT security.
Fortunately, there are…
- Every additional device increases the collective intelligence of the network.
- Patterns, Context, Identity, Behaviour can be analysed: AI, machine learning
Example: AI could be used to watch out for anomalies in behaviour and raise an alert. In this context, behaviour can be used as a proxy for identity, thereby increasing security.
John’s advice is to…
Secure the Enterprise, Not the Network
Observation: We all occupy the intersection of physical and cyberspace worlds
Suggestion: Turn all street lamps into gunshot detectors. Another example: IoT flood sensors. [Mike: Water sensors seem less innocuous than gunshot detectors].
Mike raises privacy concern citing the gunshot detector example: It resembles the sonar system used by Batman in the Dark Knight – a film, which includes a strong theme on privacy. John acknowledges the problem and suggests (a bit weakly in Mike’s opinion!) that citizens might opt in with their smartphone devices. He suggests this would needs a strong framework in law (policy, rules) and technology – as indeed it would/does. Mal Minhas (previous speaker) agrees with Mike, citing Trust as a key factor in this discussion.
John has used this opportunity to remind the audience of the incredibly important role low cost sensors will have in keeping people and infrastructure secure, though he did not appear to recognise the elephant in the room, which was obviously the potential to encroach on the citizen’s personal privacy.
Panel Session: What’s the Real Power of Data Analytics?
EmilBerthelsen (EB), Moderator
Darragh Kennedy (DK), Head of Cloud Services, Lonely Planet
Scott Amyx (SA), Managing Partner, Venture Capital
Masoud Charkhabi (MC), senior Director Quantitative Analytics, CIBC
DK: VR/AR will be huge in travel
Changes in recent years:
DK: In recent years, there have been moves to agile, devops, cloud.
SA: Organisation & change management. Execs will push back on new ways to make decisions, which will increasingly be made off the back of IoT-gathered data. SA is working on filtering data at the edge as opposed to creating “lakes of dirty water”. Data mining is looking at the past. Neural networks (as a service) and multivariate analysis can be used to make predictions.
Moderator: Is there a lack of data scientists?
DK: Yes: become one and you’ll triple your salary! Petition government to invest in this branch of science, but also use graduates and build knowledge internally.
SA: Machine learning is a set of models and an on-going process so you need continuity. But also to work with external parties.
Questioner: Raises concern about mistaking correlation for causation and bringing data science into disrepute. [Mike: great question!]
SA: “Attribution” is an aspect of the analysis that will prove indispensable.
Questioner: What problems was Lonely Planet trying to solve?
DK: Lack of scalability; responsiveness; inability to change the organisation.
Questioner: What profile for data science?
SA: A mix of skills from UX, UI, data scientists. Need to ask the question: what are we trying to get out of the data?
Moderator: How to go about finding tools?
SA: Is your organisation even ready? In any case: do things incrementally. Do MVPs. It’s not about the tool, but rather the data.
MC: Build machine learning into your day-to-day software. Put it in a position so it can learn more with every interaction.
Moderator: Themes for the next 2 years?
MC: In banking: Create scenarios i.e. alternatives and explore those alternatives. [Mike: no idea what he meant by this.]
SA: Data analytics needs to work hand-in-hand with security, privacy.
DK: Disruption has only just begun.
A bit all over the place, but I wholeheartedly agree with the observation about the danger of accumulating data and then picking out correlations, thereby establishing a correlation as opposed to a causation. There’s real potential for abuse as well as the danger of making strategic error (based on erroneously derived conclusions) here.
Talk 3: Adapting Measurement Strategies for Modern Marketing: Winning Moments in a Multi-Screen World
Jackie Wood, UK Sales Lead for Google Analytics 360 Suite
Jackie begins with some impressive stats:
- Internet moment: a gazillion interactions
- 150 interactions with phone every day
- 1 trillion connected devices. 2.5 gb of data per day.
Proposes three necessary steps in order to get a handle on this: Aggregate, Analyse, Activate.
BUT: It’s harder than you think:
- 84% of marketers surveyed don’t believe their data sources are integrated.
- 60% of marketers said that marketing measurement tools are difficult to use, thus reducing cross-team collaboration.
- 50% or marketers stated that it is difficult to give their stakeholders in different functions access to their data and insights.
[But guess what, Google Analytics 360 Suite is an end-to-end marketing suite! At this point I’ve realised we’ve moved into sales-pitch for Google product mode.]
The suite comprises 6 tools:
It’s a bit unfortunate that this talk turned into a sales pitch for Google’s newest marketing tool, even if the tool itself is potentially very interesting. If you want to know more… Google it.
Panel Discussion: Smart Cities
Emil Berthelsen (EM), Moderator
Ian Jones (IJ), Smart Cities Lead, Leeds City Council
Sean Green (SG), Service Head ICT, Customer Access & Transformation, London Borough or Tower Hamlets
Nick Wallace (NW), Government Analyst, Ovum
EB: Applications right now?
IJ: Main driver for any particular application might a theme like “make getting old better in Leeds”. This translates to choice of applications, which are weighed in light of restricted budgets.
SG: Team has created an innovation lab, which brings data and problems together. Brings in citizens to try to work out solutions. Example: Transport for kids with special needs. Innovation lab brought data to the study and helped derive a solution. Example: Use sensors to assist the elderly with independent living.
EM: What are the roadblocks being experienced?
NW: Observation: Leeds getting older, Tower Hamlets by comparison is younger. It’s easy to ignore difference between cities, which can result in failures. Politics, money are further roadblocks.
IJ, SG: Citizens are concerned with security. Privacy needs to be addressed head on. For example: Need to make the case to the public that NHS records could be used very powerfully for research.
[Mike: Solutions discussed are VERY down to Earth. Nothing particularly futuristic here. But this is probably a good thing!]
IJ: Privacy is not as black and white as many people think. E.g. share health data with GP? family? someone else? Some people may choose differently depending on situation.
SG: The Danish model is interesting. It utilises a single identifier in order to interact with all manner of government organisations. In contrast, the UK citizen requires multiple IDs (“numbers”). This makes joining the dots much harder.
IJ: Estonia is a further example of this. However, UK culture has never accepted ID cards. Suggests that acceptance of the idea of a single ID will change over time.
SG: Government is experiment with federated ID scheme, which may overcome the problem of a single ID.
IJ: Small vendors are anxious to help in turning innovation into working solutions. Larger companies are more problematic.
NW: Small, agile companies are better. [I agree!!!]
SG: Large vendors assume they will do the end-to-end solution, even if working with partners would be better overall.
It’s easy for us all to get carried away with futuristic ideas and plans. However, the panel members here remind us that a lot can be achieved just by acquiring the right data and applying it in the right place.
Talk 4: Unlocking Everything Everywhere: The Limitless Potential of People and Places
Nick Hodder: Guide Dogs (formally Google)
Contact: email@example.com … seeking partners to work with.
2 million blind people in UK
Guide dogs is an independent charity
Challenges: See slide
Websites don’t generally utilise accessibility markup. [Shame!]
Tensor Flow: Machine Learning service from Google.
Others: Theano, Torch
Cities Unlocked project phase #1: Used GPS, *Beacons* to assist people around town.
Cool video from Microsoft showing how blind developer is assisted as he negotiates various social scenarios. Heavy use of scene and facial recognition. Powerful stuff!
Nick suggests that Apple might be losing out machine learning [perhaps] because of its focus on privacy. This is what differential privacy is supposed to tackle. Apple seems to be lagging when compared to Google, Amazon et al.
Mike’s summary: IoT will clearly have big impacts on many areas of society, but surely some of the biggest beneficiaries are those citizens who can use assistance when it comes to navigating and interacting with social spaces.
Talk 5: Unlocking Big value from Data: it’s more than [IoT + Big Data + Cloud + Power]
Patrick.Bossert@networkrail.co.uk [wrote cube solve book at 12 years old], Digital Transformation Director, NetworkRail
Patrick shows colourful pic which describes the organisation.
Notes that introducing information services to an industry which hasn’t changed for 100 years is… challenging[!]
Observes [crucially] that introducing IoT etc. doesn’t guide you on how to change processes.
He then introduces a slide, which expresses the drivers of the NetworkRail process.
Key point: the process begins with the identification of a data specification i.e. the data required. Contrasts this to organisations, which begin with IoT and then try to work out what to do with the resulting data.
However, he notes, business value can also be generated in the other direction i.e. begin with IoT, although the speaker argues going left to right generates much bigger ROI. Although value can unquestionably be generated in both directions.
Another way of expressing the take-home: Use systems thinking used to guide the efforts of things which utilise IoT, Big Data etc., instead of thinking in terms of technology alone.
A bot of history: iPhone/iPad introduced 2011. Personal issue devices so people would buy into the technology. Abuse check: Manager penalised if personnel go off grid with the device. This approach built *trust*.
Next, GPS was introduced to apps. “WhereAmI” app used to locate position of fault. This assures people working on the rail system for the shortest amount of time possible – a result which has apparently saved considerable money.
Success of program overall is reflected in the enthusiasm of the personnel.
Mike’s summary: Super-smart speaker; great stuff. NetworkRail appears to be in good hands.
Talk 6: Key Trends on Enterprise Mobile Apps
Nick McQuire, CCS Insight Business Market Research Lead, Insights Enterprise Research
Opens with “Good enterprise app makes a small company look big but a bad app makes a big company look small.”
Speaker contrasts beginnings (2010-2013) with today:
2016+ Information = new battleground
- Foundation – most businesses here (55%)
- Enable – (35%)
- Holistic – (10%) large numbers of apps; process is industrialised; close with with user community; exploits measurement
[2nd (3rd?) time I’ve heard this argument today: Organisation will need to adapt in order to profit from mobile tech. Kind of a no-brainer now, no?]
- 41% or employees say mobile business apps are changing the way they work. B2B is 2nd most popular app category after games.
- 4.1 average number of business apps being used for work today
- 76% of employees have NOT asked IT to provide apps for work
- 55% of employees to NOT have a formal process to request an app
- 72% of employees to NOT have a company app store
… in summary, these stats are reflective of an wide general lack of strategy when it comes to enterprise apps. This has been RNF’s observation to-date, though we do witness this changing as we encounter more and more organisations who have explicitly mandated or hired a CTO/CIO to develop a mobile strategy. My guess is that these stats will look very different two years from now.
69% of enterprise spending is coming from the business. Even though IT spending is increasing by 5% per year.
More and more apps will be targeted to departments as opposed to being generic in nature (e.g. office).
34% develop app internally
67% have developed up to 5 apps
7 % have developed 25+ apps
A number of barriers exist to building apps (slide). No. 1 obstacle is Security.
Triangle shows app types and relative percentage of app type:
The content here was essentially the same, increasingly communicated message that we can read in many sources, albeit with more stats than most.
Talk 7: Mobile Mobility? [Yes, question mark is in the original title]
Marcia Brock, CEO * founder of The Smart World
Speaker asks us (the audience) to view a video of “Zuznow”. [No idea why, as it appears to be completely irrelevant.]
Then argues that bringing apps into the enterprise is very hard [really???]
Claims men & women interact with computers differently; furthermore, males will use their laptops more than women, who prefer mobile devices. [Evidence appears to be purely anecdotal.]
Asks some members of the audience to reveal what apps they have on their phone. She then interviews them, but doesn’t pass the microphone to them so we can’t hear the answers to her questions.
[Is this an attempt to demonstrate that the aforementioned claim is true? Does it need to be stated that a random sample of two people does NOT a compelling case make.]
When a presentation might be mistaken for a parody, you know something’s gone horribly wrong. Marcia, if you’re reading this, you need to work on your presentation skills. I mean this in the kindest possible way.
So that wraps up day #1. My brain’s fried so I guess that means I’ve been given a lot to think about.