The Journey to Managed Enterprise IoT – Part 3 – Beyond the use case

<This blog was previously published at the Atos Thought Leadership website – it was written by Philip Griffiths>

In the 1st blog in the series, an overview was shared of the journey to managed enterprise IoT, which we divide into three levels of maturity. In my second post you will learn more about the first level: Enabling the use case.

Once you have determined the IoT use case you need to ensure that the solution is secure, accurate and predictable – i.e. it delivers sustainable value – in the face of increasing quality of devices, edges, continuous data flows, and technologies. This is enabled through ‘non-functional requirements’ (NFRs) and spans both execution qualities and evolution qualities. This is going ‘beyond the use case’.

By focusing on ‘how’ the use case is delivered, instead of ‘what’, you can realize the following benefits:

  • Quality: If your products and services cannot extract value, the data, and opportunities, it is lost.
  • Risk: Complex IoT use cases can put business continuity at risk; reduce it through strong NFRs.
  • Productivity: A use case has little value if it stops working after 6 months or in peak load periods.
  • Future-proofing: Your system should be built for low cost and simple improvements / evolution.

Realizing this require you to define, discuss, document, and design your NFRs as you enable the use case and beyond it. Extending the timeline from the previous post, the following activities take around 12-24 months:

  1. Project: Design solution, install and implement IoT solution from core to edge, integrate into your existing business systems as well as ensure system security – including building ecosystem of partners.
  2. Business Platform: Scale-up and industrialize the use case to a full platform, do further roll-outs, integrate into your existing business and enhance parts of the system to delivery sustainable value.

Below is a non-exhaustive list of topics and example questions to be considered for NFRs:

NFR Topic Example Questions
Execution Availability Does your E2E use case need an operational uptime of 99.99% or 95%?
Continuity What are you E2E RTO/RPO? What are the business impact / cost of your use cases being down for 1 or 60 minutes? How does this impact backup or disaster recovery? How do you enable HA, backup or disaster recovery in a distributed architecture?
Manageability How will you update millions of devices? When and how to push new functionality? How to handle a million+ devices all calling home sick? Will you have 24×7 operations?
Interoperability  How will data be handled across different silos? Do your platforms work together?
Performance How quickly do you need to access the data? Does it need to be processed at edge?
Resilience How will the system do backups and ensure service continuity? How will you ensure high uptime of distributed architecture? How will errors be handled ‘gracefully’?
Security How will data from connected objects be trusted? How will you ensure security in new and high risk environments? How will you reduce attack surface? How will you ensure internal and external compliance, auditability as well as alignment to standards?
Usability How will you reduce the head count of managing such E2E complexity? How will you optimize latency issues to improve real-time outcomes and / or use experience?
Evolution Maintainability How will defects be corrected? Will the system and components have self-healing? How will uptime be ensured without sending people / parts onsite?
Modularity Will your system be built on principles of separate independent functionality?
Scalability Will it scale up/down to meet peak demands?

 By focusing on ‘how’ you enable value from and not just ‘what’ the use case is you will derive much greater long term value for your business. To do this, you need to define, discuss, document and design them into the use case from the start. If you are not already doing this I suggest you look at facilitating it as soon as possible.

I have spoken to many clients who have rolled out IoT solutions which are ‘the future of the businesses’. It is therefore unfortunate when they stop working effectively or if the operations team only know of problems when they are informed by the customer. It is normally at this point that they ask for expertise on going ‘beyond the use case’.  A few common outcomes are re-building the app, creating a new platform, facilitating rapid scalability or enabling an operations team with E2E monitoring; either way, it costs time and money that were not predicted in the business case.

The journey to managed enterprise IoT

<This blog was previously published at the Atos Thought Leadership website>

A lot has been said about the data-driven economy we live in and its effects on the people, processes, places and products of all businesses, but how many companies actually use their data efficiently? Every organisation should not only be driven by data or the Internet of Things (IoT – a term we will use to cover gathering, processing and extracting value from data), but also ensure that it delivers measureable outcomes into their core business processes. Few can claim to be at this level of maturity and, if you are not one of them, our overview of the journey to managed enterprise IoT can help put you on the right path.

A true data-driven strategy can deliver benefits that range from tightly integrating customer care into product usage and performance to being able to remotely monitor and improve product (or service) performance to enabling new revenue streams.

We can divide the journey towards Managed enterprise IoT into three levels of maturity (that will be further explained with their own separate blogs) including the six sequential activities shown in the picture below:

  • Enabling the use case: exploring and solving business operational problems in minimum time to value with easily understandable business impact on people, process, places and products.
     >> Benefits: Project – Business operations managing each use case
  • Beyond the use case: ensuring predictability for delivering value – i.e. high quality (from fewer errors), resiliency (higher availability and stability) and security.
     >> Benefits: Program / Portfolio – Business operations across all use cases
  • Managed Enterprise IoT: simplifying the complexity of IoT, which is now business critical, to be both agile and predictable as well as being proactive, prescriptive and automated to deliver positive and measurable benefits for your whole enterprise.
     >> Benefits: Enterprise – IoT underpins whole business

Businesses are made up of people, processes, places, and products which need to create value – the fundamentals of business that have not changed with a data-driven economy. As businesses derive more and more of their revenue from data (and IoT), they are required to resolve its inherent challenges – enabling business value, managing the quantity of things and data, dealing with the complexity of the ecosystem and reducing security and privacy risks.

This is critical to enabling benefits for companies, to their customers, shareholders and the wider data-driven economy – the outcome of Managed enterprise IoT. Rather than taking big leaps from one maturity level to another, put in place a road-map at the start of your journey to Managed enterprise IoT (particularly as forward planning will save lots of time and effort later).

The companies I work with often ask, “Should we manufacture a solution in-house or purchase it from an external supplier?” These are my high level recommendations:

  • Nothing exists in a bubble; consider your long term strategy and competencies (taking into account value and performance). Approximately, core ones should be made, marginal bought, supporting could be either – core is defined as high value and high performance, marginal as low and low respectively.
  • Do not think that you can only make or buy; you can customize a solution that fits your needs. Build a level of extraction (i.e. non-native) above the layer at which you buy; this will enable you to commoditize everything below it, to future proof and ensure cost reductions and quality improvements.
  • Open source gives you access to a larger pool of innovation and can be a low cost way to experiment (enterprise versions offer greater predictability).

In the next post, I will explain the 1st level of maturity of this journey towards Managed enterprise IoT: ‘Enabling the use case’.


I would like to add a special thanks to Philip Griffiths (@ThePGriffiths). Philip was until recently, the strategic partner manager for the ATOS IoT practice and took the initiative to write this blog-series you are reading now.

 

IPv6 – your next cash cow?

For anybody looking at the next big thing, the new 'killer app' or the new gold, I recommend to read a white paper by the Atos Scientific Community called "IPv6: How Soon is Now?".

The paper explains very well the problem with the way the internet is currently working. It points out that we have a serious issue, a 'time-bomb', with the way that devices (computers, networking components and other IT stuff) are connected with each other using this old IPv4 technology. The paper further explains why, in spite of all kinds of intermediate technologies, we need to adopt a new technology, called IPv6, and we need to do that very quickly.

"To sustain the expected Internet growth, there is no adequate alternative to adopting IPv6."

Furthermore you will read in the paper that we will be running into real problems if we do not make that change and unfortunately the change is happening much too slow.

"Unfortunately statistics from Google paint a (…) picture with less than 1% of total users being IPv6 clients"

This might sound awfully boring and a field of play for the technology wizards in your organizations – this is not for you right? But wait, because halfway through the paper, the authors start explaining that the benefit of this new technology is in the way it can support all possible technical devices (including cars, phones, traffic lights, wind mills, televisions, your shoes and wrist watch, medical devices and almost anything) can become connected – can talk with each other – when we switch to IPv6.

"(…) that IPv6 can now be used on virtually any communicating object, from servers to small sensors, regardless of the underlying (…) network technology."

I think this changes everything; it opens up a whole new world of play for consumers and manufacturers, for service providers and retailers; to create new businesses, to open up new markets and create new ways of making money.

"The IPv6 "Killer App" is likely to be the enablement of the Internet of Things (IoT)"

Based on this you would be stupid to not support this move to IPv6; it will be the engine that allows your business to innovate and grow; your IT landscape will increase thousand fold and you can bring any type of information, sensor or other device into your business platform. That is cool and exciting.

But it will not be easy.

"Although many people think that a migration to IPv6 is primarily a networking issue, the truth is that all IT organizations across server, network, storage and application domains must be equally trained to contribute to both the planning and execution."

The authors explain in quite some detail that you will need to overcome technical hurdles (IP Space Management, IP Address Provisioning, IPv6 to IPv4 interoperability, Application IPv6 readiness and Security Challenges) as well as business challenges (Coordination across silos and companies, Timing issues on what to do first and governance to establish End-to-end responsibility).

"We predict a tipping point when there will be more IPv6-connected users and devices, and therefore opportunity, than the IPv4 landscape provides today."

So, want to grow your business, do the strategically right thing and set yourself up for business growth, agility and all the other stuff you need and like? Migrate to IPv6 now.


This blog post was previously published at http://blog.atos.net/blog/2013/09/03/watch-this-space-lucky-7-avoiding-information-overload/

The Data ‘Explosion’ is real

In a recently published research called Ascent Journey 2016, the Atos Scientific Community considers the massive growth in data and storage as an important trend in IT.

Whilst the concept of Big Data has been around for a number of years and is relatively well understood, it is now becoming clear that everything we do is leaving a trail of data that can be analyzed and used.

Examples include the payments we make on a credit card, the books we read on an e-reader and our energy use by driving an electric car. This will lead to a new era of Total Data that, in turn, will lead to new business models, services and economic growth.

We don't yet understand all the implications of this – for businesses and society – but organizations that are able to harness and make sense of the vast quantities of heterogeneous data from disparate sources will gain valuable insights into market trends and opportunities.

An 'Ecosystem' of new management tools is taking shape, covering the various layers of the data stack in the enterprise and capable of delivering a 'Total Data' approach.

The technology that supports the Information Management Lifecycle in the enterprise is going through a profound change, due to the emergence of new solutions, many from open source background (NoSQL databases, Hadoop, analytical tools like R, visualization tools). To enable the 'Total Data' environment, the new technologies need to connect into and partly replace traditional technologies.

  • In some scenarios, data must be obtained, processed and correlated with insights being derived and actions initiated as close to real time as possible.

Yesterday's data is not interesting unless it helps predict tomorrow. Yesterday's traffic report isn't helpful in plotting a journey today unless it is known to represent today's pattern as well, and combined with other data can improve future congestion. Pattern Based Strategy enables huge amounts of historical data to be analyzed for previously invisible patterns. These patterns give us the power to start predicting what is likely to happen in the future, so we can plan and improve, both in real-time and in non-real time scenario-planning. For example, real time predictive analysis will plot the route for transporting donor organs across a city safely and quickly, continuously adapting the route to changes in the traffic patterns as they are happening. Another example is a country to make compliance recommendations (and potentially becoming a legal requirement) to companies for maintenance regimes for their infrastructures or industry plants using analytics on historic data and thus establishing an automated "what are the lessons learned" process.

  • Everything will be digital and everything will be connected.

Everything will be captured; "your life is becoming a video" – you can even replay your actions, thoughts and analyse in various forms and for multiple purposes (see http://quantifiedself.com/ for example); this is not only becoming possible for peoples life's, but anything that can be measured can be tracked, traced and put in a digital context for analysis. The ability of businesses to process this wealth of information is still unclear. What these developments – and others related such as 3-D printers and cognitive computers that will be able to replicate smell and touch for their users – mean for society, laws and concepts such as individual privacy need to be reassessed and will prove a huge challenge to governments, businesses and individuals in the 21st Century; for example long-established laws and concepts such as individual privacy need to be reassessed.

  • After an initial confusion phase, traditional and 'Big Data' orientated approaches to analytics will converge in a unified 'Total Data' platform.

Big data relies on its sister technologies of optimized IT networks, rapid mobilization communication tools and cloud computing. Data Analytics as a Service could emerge from a combination of Big Data, Pattern Based Strategy and Cloud technologies. Business performance can improve in areas such as increased forecasting and enhanced automation capabilities and buildt new business propositions upon the discoveries they can do using Total Data as a source of undiscovered information.

[This blog post is a rewrite of http://blog.atos.net/blog/2013/02/22/the-data-explosion-is-real/?preview=true&preview_id=1555&preview_nonce=6df2f23c80 ]

How big is your robot?

What do you get when you combine cloud computing, social networking, big data and modern day engineering? You get a kick-ass robot. This was my first thought when I finished reading a published whitepaper by the Atos scientific community on the topic of robots.

Central in the paper is the question: “Where is the mind of the future robot?”, and by outlining the concept of a robot that can utilize everything that is available in cyberspace you may find it difficult to answer that question.

Today it is hard to predict where on earth all of the data about you is stored in the cloud and we have never been able to communicate more easily. It is easy to see that robots will be everywhere, able to utilize all available information.
This will lead to a new class in robot persona’s and capabilities.

Once the robot is part of a social network, it could virtually interact with humans as well and thus start truly mimicking human behavior.


When I was (much) younger we had a program on our home computer that was called ‘Eliza’. This program would behave as an electronic psychiatrist. It had some limited learning capabilities and some clever language skills to ‘trick’ you in having an actual conversation.

If you would type things like “I hate talking to a computer”, Eliza would answer with “Hate seems to be important to you, can you explain that?”

If we now multiply the capabilities of this ‘Eliza’ by a thousand or more (using cloud computing scalability) and bring in the analytics of all of your ‘likes’ or ‘diggs’ or even the behaviour of your friends, combined with knowledge about your locations and multiply that by analysing all the things you did 5 years ago, 10 years ago and today …. Well I think you get the picture.

The more a future robot knows or has access to, the more it will be able to fulfil his role in supporting us. This may not sit well with everybody, but if we utilize this capability in a clever way, I believe we can benefit.

Especially if we also take into account that a robot can take different forms, could exist virtually or maybe even be in multiple locations at the same time, with access to the right information and computing power to use that to our benefit. The whitepaper describes some of these scenarios and puts it in the perspective of the role of IT providers and systems integrators.

Based on my reading of the whitepaper I was thinking that maybe the statement ‘I cannot be in two places at the same time’ will soon become a thing of the past.



[This blog post is a repost of http://blog.atos.net/blog/2012/11/26/watch-this-space-how-big-is-your-robot/ ]