SOGETI UK BLOG

Supercomputing (aka High Performance Computing, aka HPC) is no longer a tool for engineers and scientists, but has become a part of the new infrastructure for more and more companies.

One example is financial services that use HPC, linked with big data analysis to create new solutions for their employees and customers. This alliance coins a new term HPDA for High Performance Data Analysis.

HPDA, as defined by IDC, is composed of four segments:

  • Fraud detection: This segment focuses on the identification of suspicious or potentially suspicious behavior and their causes by using graphical analysis, semantic analysis, or other high performance analytical techniques.
  • Marketing: This segment covers the use of HPDA to promote products or services, usually using complex algorithms to discern the demographic characteristics of potential customers, their preferences and consumption patterns.
  • Business intelligence: This segment uses HPDA to identify business opportunities for increasing competitiveness of enterprises by a better understanding of their activity, their competitors, and the dynamics of the markets in which they operate.
  • Commercial HPDA: This segment includes all commercial HPDA initiatives other than the three that come to be described.

In the near future, a new segment will appear in the field of personal medicine. Application will focus on the analysis of the medical results and patient history for purposes of diagnosis and treatment planning compared with data coming from overall population. While a patient will be in his doctor’s office, the latter will be able to correlate the millions of archived records of patients suffering from similar symptoms to get relevant results. The doctor should thus take into account the effectiveness of different treatments for patients with similar pathological profile.

Another field for HPC and business is the 3D visualization, on your own device or using cloud infrastructure. One example of new application is Remote 3D visualization for real estate service:” Bouygues Immobilier” is the first customer of the young enterprise “MyCloud 3D”, which model 2 Green Office building. This solution was created for purposes of promotion and real estate marketing to help selling premises all over the world, offering virtual visits and re configuration on the fly

However, there are areas for improvements:

  • Software remains a major obstacle. The improvement of management and administration of the HPC resources software are essential and there is a persistent gap in parallelized algorithms for the majority of users.
  • Clusters are still difficult to deploy, use and manage.
  • Storage and data management are more and more bottlenecks because of the significant increase in the volume of data.

Tomorrow (but sometimes, tomorrow is already real!), HPC capabilities will be embedded in small appliances: Machine Intelligence technologies like IBM Watson sits now in a pizza box and tomorrow in your smartphone or tablet.

References

IDC’s ISC High Performance, The HPC event, July 2015

 

Philippe Andre AUTHOR:
Philippe André is an expert within Business and IS architecture, Service Architecture, System modelling and Soil science. Philippe is a Certified Enterprise Architect (L4) and TOGAF9 certified. Philippe’s mission is to help clients to make the best decision as far as business and IT alignment is concerned. He works as a link between architecture and design team, making sure that architecture decisions and directions are applied on the field.

Posted in: Big data, Business Intelligence, Cloud, Data structure, Digital, High Performance Analytics, Infrastructure, Innovation, Marketing, Social Aspects, Social media, Software Development      
Comments: 0
Tags: , , , , , , , , ,

 

google car

As we come to the end of 2015 today, let’s look at the Top Blog of 2015!

At the last CES held in Las Vegas, Google introduced solutions that could revolutionize all industries related to mobility.

Google has a huge lead over other automakers with its Google car. First, the reliability of a driving-related software / technology can only be established after it covers millions of kilometers. And, in 2013, Google had already covered a million miles on American roads with its fleet of 25 autonomous cars, much before all other constructors. The second key advantage and differentiator, in addition to embedded sensors, is Google’s ability to provide a system of mapping in 3D, allowing a very accurate identification of car position (with GPS 2D, the uncertainty is 7 meters, which is insufficient in terms of secure driving). Google Maps has, now, mapped one third of the paved roads in the world, against about 7% for Nokia Here, its first competitor! Finally, facing Google, automotive industry advances in a dispersed order.

If Google has the ambition to propose the Google car on a free sharing basis (a reasonable assumption), this will have a very big impact on the entire automotive industry. The development of the autonomous car on free sharing basis will affect the sale of vehicles (and therefore all actors in the chain, up to insurers). It will also represent a new alternative solution to traditional transportation (bus, tramway, subway, small distance trains). The automotive sector provides 10 million jobs in Europe, mainly in distribution networks, repairing, insurance and infrastructure. So, autonomous cars might lead to millions of job losses!

The passage to the autonomous car, following the Google business model, therefore, represents a full paradigm shift for the entire sector. And time is running out for the automotive industry to organize itself, because Google has already announced that it will be able to industrialize this car before 2020, i.e. only five years from now!

However, it will bring undeniable benefits to the society as a whole, by allowing the transition from an economy based on possession to an economy based on the usage. This will be more favorable to sustainable development, by reducing road accidents, by ensuring the autonomy of the elderly and disabled, as well as by freeing valuable time that’s devoted to the conduct, today.

After all, a medal always has two sides!

References:

  • Franck Casenave, Stop Google, Pearsonn edition, 2014

 

Philippe Andre AUTHOR:
Philippe André is an expert within Business and IS architecture, Service Architecture, System modelling and Soil science. Philippe is a Certified Enterprise Architect (L4) and TOGAF9 certified. Philippe’s mission is to help clients to make the best decision as far as business and IT alignment is concerned. He works as a link between architecture and design team, making sure that architecture decisions and directions are applied on the field.

Posted in: Human Interaction Testing, Infrastructure, Innovation, integration tests, Mobility, Quality Assurance, Research, Software Development, Test Driven Development, Testing and innovation, Transformation      
Comments: 0
Tags: , , , , ,

 

Iot (Internet of Things) is full of initiatives that could lead to incompatible or antagonism solutions and standards. The last IDF (Intel developer Forum) this year provided a 360° vision of what could be a future IoT ecosystem. IoT initiatives could be categorized in 3 different development streams:

Sensification

Some examples of sensification are already up and running today:

  • Wake and voice : The machine is constantly listening (even in stand-by mode) to its environment thanks to audio technology embedded into the micro-processor. This technology is included for example into Cortana and windows 10 from Microsoft.
  • Realsense is used for :
    • Drone and other type of machine avoiding obstacles along their route
    • You can capture in real time your environment in 3D with your smartphone (Tango project for Google)
    • In some hostels, you are already welcomed by robot (butler project of Savioke)

It is interesting to notice that lots of initiatives leverage open source technologies (e.g. Open source Robotics foundation)

Smart and connected world

How to personalize the world to the user context and needs?

  • Look to yourself in a mirror and change colors of a shirt without changing of physical shirt (i.e. this is the “mirror” that changes the color for you when you choose to do so) like the virtual shopping project by Memomi

Solutions are developed based on IoT developers frameworks linked with big data technologies that offers reusable and packaged services ready to use and to integrate like:

  • Turn data into insight
  • Gateways to any protocols
  • Connect things and devices
  • Intelligence at the edge (data filtering, data analysis and sensors monitoring done by the sensor or centralized from a datacenter)

Extension of you

Smart and connected devices allow the technology to be part of you and extend your possibilities like:

  • The wearable products (connected watch for example)
  • The Curie project can acquire automatically lots of data in real time. It is a micro computer (button sized!) with embedded Bluetooth technology. Intel has developed a BMX sensor in order to capture in real time the bike spin, max flips per second, etc… It is used in athlete enhancement programs.
Security for IoT

Security is a key (even a killer) concern for IoT wide adoption. Developing and proposing consistent platform and component to ensure appropriate level of security is an important challenge today.

This year, Intel presents the enhanced privacy identification component and is working on a security reference model for IoT security.

Capabilities of these solutions embed security measure in a specific component that interact with the device. In wearable world, the sensor (e.g. carried by a BMX) is activated by a wearable product which is personal to the user. When the user is far away the sensor, this one is inactive; as soon as the wearable product is near the sensor, this one is activated. If the wearable product is physically separated from the user, it does not work anymore, so the sensor cannot be activated anymore.

Software and hardware interaction

IoT Platform of the future will be software and hardware together. And from an industrial ecosystem standpoint, interaction between software and hardware vendors must exist to ensure smooth integration. And standards could help in this area.

One special topic is the storage, in the machines itself and in the data center that receive and compute the data from multiple devices. In the near future, new storage technologies (like 3D Xpoint from Intel) will be included into the processor (e.g. 3D Xpoint technology included into Xeon processors). New software solution will be developed to leverage these new capabilities

Promoting co-operation and standardization

The Intel vision provides solutions (hardware and software) to the IoT developer community that covers all streams. In order to develop interoperability at hardware and software levels, such industrial initiatives must group different actors of IoT world, forming consortiums (including Open source communities) , otherwise the IoT ecosystem will be a mess, even a dream.

References

Intel Developer Forum 2015 presentations and white papers

 

Philippe Andre AUTHOR:
Philippe André is an expert within Business and IS architecture, Service Architecture, System modelling and Soil science. Philippe is a Certified Enterprise Architect (L4) and TOGAF9 certified. Philippe’s mission is to help clients to make the best decision as far as business and IT alignment is concerned. He works as a link between architecture and design team, making sure that architecture decisions and directions are applied on the field.

Posted in: Big data, Digital strategy, Internet of Things, IT Security, IT strategy, Quality Assurance, Research, Security, Smart, Software Development, Transformation, User Experience, Wearable technology      
Comments: 0
Tags: , , , , , , , , , ,

 

(This article is a review and summary of different white papers, books and courses referenced in the bibliography section, focusing on business models and some examples of them.)

The Internet of Things (IoT) represents the vision that every object and location in the physical world can become part of the Internet: Objects and locations are generally equipped with network, storage and computing capacities and so they become smart objects that can take in information about their environment and communicate with the Internet and other smart objects. So, smart things are hybrids, composed of  elements from both the physical and digital worlds (see fig 1. from [1])

Image 1

Figure 1 : Value‐creation Layers in an Internet of Things Application (in [1] Fleisch, E., Weinberger, M., Wortmann, F)

The key layer connecting both worlds is the connectivity layer. The physical world is not just about hardware, but also more and more software nowadays (think cars). However, a physical thing becomes “smart” when it connects to the digital world. The layers 2,3 and 4 allows us to invent and propose to individuals (customers but also citizens) new services (digital services of layer 5). One important fact is that layers 1 through 5 cannot be created independently of each other. That is why the arrows connecting them are bi‐directional in fig 1. An IoT solution with value is usually not the simple addition of layers, but rather, an integration extending into the physical level. How the hardware is built, for instance, is increasingly influenced by the subsequent digital levels and on the other hand, software which compose the digital levels and must be designed to fit the physical levels.

The connection between the two worlds allows new business models to emerge across manufacturing industries (i.e. industry that designs and builds physical world products). Here are some examples:

  • Canary (www.canary.is) or Netatmo (www.netatmo.com) companies offer a smart home alarm system that includes a variety of sensors, from temperature or movement sensors, up to a camera. The basic function of monitoring rooms during the resident’s absence and sending a message to a smart phone app in the event of anomalies, is included free of charge in the price of the system. Other services are proposed with additional cost. It is the physical ‘freemium’ business model.
  • You have a car and you want to travel abroad, where not all risks are covered by your insurance. In this case, if your automobile’s performance can be configured using software and the vehicle is a node on the Internet, you can purchase from third party the right additional mini-insurance policy just for this journey. It is the Digital add-on business model.
  • By pointing a smart phone at a product, an Internet website opens where that same product – including replacement parts, accessories, and consumables – can be purchased. Amazon already offers this type of service for products with a bar code that are carried by the e-tailer. In this case, the product itself becomes a point-of-sale. It is the “product as point-of-sale” business model.
  • Your heating system orders itself oil refills as soon as a certain level of liquid is noted in the oil tank. You have nothing to do! The Thing has the ability to independently place orders on the Internet. The idea of self‐service no longer refers only to the customer; now things can serve themselves too. It is the Object self-service business model.
  • Brother (www.brother.com), offers leases for laser printers, for example, without any base leasing rate – only the pages that are actually printed are invoiced. In the past, the technology required to monitor remote usage was complicated and relatively expensive but as the IoT expands itself quickly, the costs required diminish, making this pay-per-use model very effective and attractive for customers. It is the remote usage and monitoring business model.
  • Another powerful idea is collecting, processing, and selling for a fee the sensor data from one sub-section to other sub-sections of the fog and Cloud computing infrastructure and ecosystem. The measurement data from the physical world are no longer vertically integrated, collected, stored and processed for just one application but instead for a broad array of potential applications. In this Sensor as a Service business model vision, value is created by making sense of data, not from the physical product itself, which will be usually inexpensive (at least compared with existing solutions until now). Fig.2, from the 2014 Vision Mobile report (reference [4]) illustrate this point.

Image 2

Figure 2:Value in IoT is created by making sense of data (report: Breaking free from internet and things – vision mobile 2014 – CC-BY-ND)

Implementing these business models is not trivial. Issues arise from the key different characteristic of the physical world (i.e. product) and digital world (i.e. service)

  • First, services differ fundamentally from products because they cannot be stored and as a rule they are provided at the customer’s site while collaborating (there are some interactions human or IT related in the delivery of the service), and they are generally paid for, in many smaller amounts, spread out over time.
  • Second, the difference between physical and digital products are particularly noticeable in product development. In the digital world, agile development processes are the norm today. When a bug can be repaired with an update at almost no cost, even in an installed base counting millions of instances, speed, early customer contact, and aesthetics are of utmost importance in development. In the hardware business, however, and in the world of embedded computing as well (i.e. mix of hardware and firmware/software), an error in a product that has already been sold usually results in an extremely costly, image‐damaging recall action (remember car recall action because of mechanics or software defects)

However, nowadays, digitalization of hardware function is becoming more and more important and advanced (see Software defined X technologies in IT infrastructure –e.g. Software defined Network, Software defined storage, etc…) and this trend will for sure help to solve the gap between physical and digital cultures.

Traditional business models put emphasis on products, technology and specific verticals. That makes sense when customers are known and their needs are well understood. The IoT is different. The demand for billions of connected devices will come from people using services and apps that make sense of the data that those devices generate.

The future leaders of the IoT will win by building community of entrepreneurs around of their products and services, which means put in place technology community platform (i.e. a system that can be adapted to countless needs and niches, see Marc Andreessen – founder of Mosaic and Netscape – definition of a platform). To be successful you will have to master both technology and the human art of managing ecosystems.

It is the key success factor to build the new business models fostered by the IoT.

Bibliography:

[1] Fleisch, E., Weinberger, M., Wortmann, F.: Business Models and the Internet of Things,

Whitepaper of the Bosch Internet of Things and Services Lab, a Cooperation of HSG and

Bosch (2014). www.iot-lab.ch

[2] Fischer et al. (2014); Fischer, Thomas; Gebauer, Heiko; Fleisch, Elgar: Service Business Development:

Strategies for Value Creation in Manufacturing Firms, Cambridge University Press, 2014

[3] Fleisch 2010: What is the Internet of Things? An Economic Perspective, Auto‐ID Labs White Paper WPBIZAPP‐053, ETZ Zürich & University of St. Gallen, January 2010,

[4] Schuermans &Vakulenko (2014); Schuermans, Stijn; Vakulenko, Michael: IOT: Breaking Free From Internet and Things, How communities and data will shape the future of IoT in way’s we can’t imagine.

VisionMobile Report, 2014

[5] Fog Computing and Its Role in the Internet of Things, Flavio Bonomi, Rodolfo Milito, Jiang Zhu, Sateesh Addepalli, Cisco Systems Inc, 2012

[6] The Fog Computing Paradigm: Scenarios and Security Issues; Ivan Stojmenovic, Sheng Wen; Proceedings of the 2014 Federated Conference on Computer Science and Information Systems; ACSIS Vol.2, 2014

[7] Fog Networks and the Internet of Things; MooC learning; University of Princeton; Mung Chiang and al., 2015; https://fr.coursera.org/course/fog

Philippe Andre AUTHOR:
Philippe André is an expert within Business and IS architecture, Service Architecture, System modelling and Soil science. Philippe is a Certified Enterprise Architect (L4) and TOGAF9 certified. Philippe’s mission is to help clients to make the best decision as far as business and IT alignment is concerned. He works as a link between architecture and design team, making sure that architecture decisions and directions are applied on the field.

Posted in: Agile, Cloud, Data structure, Digital, Digital strategy, e-Commerce, Human Interaction Testing, Innovation, Internet of Things, Marketing, Open Innovation, Software Development      
Comments: 0
Tags: , , , , , , , , , , , , , , , , , , , , ,

 

shutterstock_184875824-1728x800_cAfter a few years of definition, proof of concept and operational validation, virtualisation and hyper convergence technologies are now mature enough to invest in the data center.

It is now the passage to the SDx (Software Defined X technologies) which characterises them, to gain flexibility, agility and competitiveness.

Welcome to the dawn of a new infrastructure world, governed by the software!

Virtualise the network: remove the inconsistency

The potential of a Software Defined Network (SDN) lies in its ability to facilitate changes in the organisation and support different processes, while the control of the resource is managed by people outside the network itself or even more, external to the enterprise itself. The goal is to simplify and ease the operation that generates many new possibilities of optimised usage and configuration of the network.

The first idea is to remove, as far as possible, the material constraints in all processes, by putting the intelligence of the exchanges and of the organisation at the level of the software and not anymore inside the hardware. By doing it this way, we can use standard and commoditised hardware instead of advanced components to implement and run the network. Putting as much as possible in the software level means that security MUST be embedded from the beginning in the design as a mandatory and critical requirement!

There is one vision to SDN, but different implementations like Openstack and OpenFlow from ONF (Open Network Foundation) or VMware approach exist together (see references section hereafter).

A bunch of benefits

There are a bunch of benefits to the SDx approach:

  1. Cost optimisation, i.e. the reduction of the cost of upgrade of the hardware against costs generated by errors of development – it is otherwise simple to modify software than hardware often returned to factory to be updated
  2. Agility, i.e. reduce the complexity of the network through implementation of intelligence at the software level and accelerate the implementation of new services within the network – this can be simplified by the deployment of the software layer above the hardware one
  3. Standardization, i.e. the arrival of industry standards that will allow the organization to play to competition
  4. Flexibility, i.e. once the servers and adjacent resources are virtualised, their creation and assignment (internally and externally) to meet the requirements of the company is greatly facilitated, with operational and considerable delayed gains
  5. Control, i.e. technology best practices and state-of-the-art provide full visibility and full control on IT resource usage with the capability and ease-of-use of the software we can configure easily – at least more easily than hardware!

References:

[1] DataCore’s Fifth Annual State of Software-Defined Storage Survey, http://www.datacore.com/blog/posts/2015/05/12/now-available-datacore-s-fifth-annual-state-of-software-defined-storage-survey

[2] VMware EVO:RAIL, https://www.vmware.com/products/evorail

[3] Nutanix Virtual Computing Platform, http://go.nutanix.com/rs/nutanix/images/Datasheet_Official.pdf

[4] HP ConvergedSystem 200-HC StoreVirtual System, http://www8.hp.com/us/en/products/data-storage/hyper-converged.html

[5] Scale computing, http://www.scalecomputing.com/hyperconvergence/convergence-in-the-datacenter/

[6] Openstack foundation https://www.openstack.org/

[7] Openflow foundation https://www.opennetworking.org/sdn-resources/openflow

Related Posts:

  1. Cloud and Open source go hand in hand
  2. Hyper-Convergence in Data Center: Promises and Reality
  3. How hyperconverged infrastructure convinced me?
  4. Big Data & Analytics in 2015: Development of open innovation platform to drive business results

 

Philippe Andre AUTHOR:
Philippe André is an expert within Business and IS architecture, Service Architecture, System modelling and Soil science. Philippe is a Certified Enterprise Architect (L4) and TOGAF9 certified. Philippe’s mission is to help clients to make the best decision as far as business and IT alignment is concerned. He works as a link between architecture and design team, making sure that architecture decisions and directions are applied on the field.

Posted in: Software Development, Technical Testing, Technology Outlook, Testing and innovation, Transformation, User Experience, Virtualisation      
Comments: 0
Tags: , , , , , , , , , ,