Digital Marketplaces Using Novel Infrastructure Models.

http://sne.science.uva.nl/ http://uva.nl/

Session at 2018 GLOBAL SUMMIT Washington on Wednesday May 9, 10h45-11h45.

Name:                 Cees de Laat
Affiliation:          University of Amsterdam
Email:                 C.T.A.M.deLaat=>uva.nl

Session Abstract:

Next to data intensive science, industry is also facing exponentially growing rates of data generated by sensors placed in equipment, vehicles, manufacturing plants, refineries, etc.

Considering such use-cases, whilst bringing them alongside solutions serving science, provides an excellent opportunity to learn from each other an increase the wealth of requirements.
An important challenge both science and industry faces concerns how infrastructures should consider data-ownership rights whilst enabling the creation of value from sharing data. Organizations, that normally compete with each other, increasingly find the need to share data to accomplish common goals no single organization can create on its own. Such data sharing is hindered by the absence of trust among those organizations that shared data will only be used for the purpose for which it is shared but not for other (competitive, litigation,etc.) purposes. Examples of such use cases can be found in science (life sciences), industry (preventive maintenance, health), and society (smart city, decision support on crowd management) projects.

In this session, we will introduce the concept of a digital marketplace that enables an ecosystem driven by agreements and compliance arranging exchange of data. Such ecosystem supports enforcement features that allow organizations to manage and control risk, and therefore trust, when providing data for commonly used applications or application development. The concepts introduced here build upon internet exchange models and raise the peering models to the data layer.

This session will have several speakers presenting the needs from the Industry and the Science domain. Then, we will have speakers present the architectural and methodological challenges of building such infrastructure to support Digital Marketplaces. The session will include demonstrations of Data Marketplace principles for preventive maintenance in the Airline Industry. Here terabytes of data must be shared in secure and trusted environments allowing experts to monitor health of Aircraft systems like its engines. Potentially we will also demonstrate Container Networks that create secure overlay infrastructures to invoke and enforce data policy.
Much of this work can be seen at http://sc.delaat.net and http://delaat.net/dl4ld.


All slides shown in the session are concatenated in one pdf and can be found here: NOT YET ;-)

After the session we will present small demo's about our Docker approach.

Background information

  • Report from NWO/STW Workshop “ICT with Industry 2016” Lorenz Centre Leiden, Nov. 7-11th 2016; Prof. dr. Tom M. van Engers (UvA), Prof. dr. Robert Meijer (UvA, TNO), Dr. ing. Leon Gommans (Air France KLM Group ICT Technology Office R&D, UvA), Dr. Kees Nieuwenhuis (Thales Nederland B.V., CTO Office), "Trusted Big Data Sharing for Aircraft MRO using a Secure Digital Market Place mechanism."
  • Trusted Big Data Sharing; Researching alliances and infrastructure models across multiple autonomous organizations presentation.
    • Leon Gommans, Ameneh Deljoo, Ralph Koning, Ben de Graaff, Tristan Suerink, Gerben van Malenstein, Axel Berg, Erik Huizer, Rob Meijer, Tom van Engers, Cees de Laat.
      This effort researches the concerns many organizations have that prevents them from sharing their Big Data Assets considering the associated risks. We show how some of these concerns can be addressed by creating an alliance organizing and maintaining trust amongst members of a group that see a particular common benefit. We also consider a number of Big Data Sharing infrastructure models, implementing alliance rules using a common digital marketplace to administer and enforce them.
  •  DockerMon: Bring your own container - demo @ SC16:
    • Lukasz Makowski, Daniel Cabaca Romao, Cees de Laat, Paola Grosso.
      Our research on container-based remote data processing investigates the applicability of container technologies for sharing of (scientific) data. We focus in particular on the analysis of the challenges and requirements posed to the overlay networks interconnecting the containers. Scientific datasets are usually made publicly available, however, the data cannot always leave the organization premises.  Moreover, on-site data processing can be challenging because of incompatibility of systems, lack of manpower or the volume of the dataset itself. We develop a proof-of-concept employing containers performing data retrieval and computation networked with VXLAN overlay. The user is given the ability to create containers equipped with the chosen set of functions. Where each function is capable of returning a different subset of information. Next, the copies of the container are concurrently executed at the different locations holding diverse datasets. The output of such execution is the data found by a particular function. Finally, the multiple results are correlated and returned to the user. Our SC16 demo is a gamification of the remote dataset processing architecture. The selection of container functions is constrained by the budget i.e. each function costs a certain amount of money. Additionally, the ability to run the created container at a selected location also requires a fee. The user picks different search functions, represented as tools, to find animals in the remote datasets. Lastly, correlating found animals according to the correlation method of choice. See:
  • Unlocking the Data Economy via Digital Marketplaces; Researching governance and infrastructure patterns in airline context. - demo @ SC17:
    • Leon Gommans, Ameneh Deljoo, Joseph Hill, Paola Grosso, Lukasz Makowski, Gerben van Malenstein, Dirk van den Herik, Wouter Kalfsbeek, Teresa Bartelds, Axel Berg, Cees de Laat, Robert Meijer, Tom van Engers.
      Data sharing and digital collaboration in logistics is important for increasing efficiency, lowering costs and lowering pressure on infrastructure and environment. Furthermore, digital collaboration makes the creation of new logistic concepts possible, leading to new business opportunities and providing solutions for challenges like visibility of goods, synchronization of planning among partners and bundling of capacity. A data infrastructure must be established where data sharing among logistic partners is easy and robust and can be set up in an ad-hoc fashion. Agreements for data sharing between partners are secured in the infrastructure, data owners have full control over who has access to what data and for what purpose and a service industry has arisen that offers logistic and infrastructure services on the infrastructure. The infrastructure will maximize business value, comply with various legal requirements whilst allowing partner autonomy, and enhance existing data sharing and storage facilities. See: 
  • Light Paths and Data Transfer Nodes for  Aircraft Maintenance; Data Transfer Node (DTN) Workflows. - demo @ SC17:
    • Joseph Hill, Gerbenvan Malenstein, Leon Gommans, Cees de Laat, Paola Grosso.
      Air France-KLM uses a 100 Gbit/s link, connected to Netherlight, to research an aircraft maintenance industry use case. Via this open exchange, Data Transfer Nodes (DTNs) of Air France-KLM in the Netherlands and iCAIR – present in Chicago at StarLight – connect to each other using light paths over their links. In this demonstration, users at SC’17 in Denver will experience the difference in transfer rates with and without using DTNs. See: