[IOT] Google & "Open Web of Things"

Peter Waher Peter.Waher at clayster.com
Fri Dec 19 21:10:22 UTC 2014


Google has asked for proposals regarding the "Open Web of Things", and they are interested in funding interesting projects (500.000-800.000 USD is it is a group of researchers at a university, or 50.000-150.000 USD if it is a private investigator). If you have not seen it:

They have five categories:

1)      Human-Computer Interaction

2)      Privacy & Security

3)      Systems & Protocols

Since we are strong within these areas, we can offer help to mount such projects. (The grants go only to the university involved.) What we can do:

*         Help with project ideas.

*         Assist in the technical description of the project.

*         Provide technology and a platform for rapid application development for IoT.

*         Act as mentors.

*         Help with evaluating the project.

What we cannot do is help with the actual development, nor in the description of the results, or hardware for the project. (Such things need to be covered by the grant.)

The proposals (which can be various) have to be sent before the 21st of January 2015. Here comes a few proposals of projects that can be done, and that fall within what they are looking for:
Securing the Life Cycle for things and provisioning for Internet of Things using XMPP
The majority of solutions for the IoT are manually configured. Very few solutions exist where the end user can buy a product designed for massive use in a retail store, connect it, and then use it on the Internet in a secure manner.

We have made a proposal of a distributed architecture that takes the entire life cycle of an IoT product into account, designed for massive scale, from production, sales, installation, creation of identities, configuration, authorization, change of ownership and decommissioning, in a secure and scalable way, without losing real-time communication capabilities. The architecture also supports distributed transactions, protects information (the owners of the things are also owners of all its data and control who can access it and do what with it.)

You can find more information about the control of a things life cycle here:

More information about provisioning of things (who can access what things, and read/control what data):

More details:

The project could include:

*         Implementation of the indicated protocols for use in things. (Open source code exist on GitHub, for reference: https://github.com/Clayster/Learning-IoT-XMPP)

*         Implementation of the indicated protocols for servers, or use of existing servers. (The thingk.me platform already supports them)

*         Define "normal use"

*         Measure efficiency of different use cases.

*         Simulate large volumes of things, and different distributions of servers.

*         Security analysis.

*         Proposal of extensions.

Extension of the Semantic Web to Peer-to-Peer-like Sensor Networks based on XMPP
One of the problems with the Semantic Web, or the Web of Things, is the problem of security and distributed transactions using the HTTP protocol. In the traditional web, the web servers are centralized, and security is managed on the same web servers. For the Web of Things, each sensor has the possibility to become a small web server. How do you manage security in a similar distributed environment? HTTP does not have any good solutions by itself, except it supports the use of directory services (which are designed for use within intranets and not on the global Internet). The same happens for the Semantic Web or with SPARQL. Another problem is how to bypass firewalls. Normally, sensors are installed in protected environments, behind firewalls. This This prohibits access to the sensor directly from outside of the firewall, unless holes are punched in the firewall.

To resolve this, most solutions store all information centrally. This later permits the management of security to that centralized location. But, by doing this, you lose the local control and the real-time communication with the devices, something that prohibits many different types of applications, and limits the scalability of the implemented solutions. Furthermore, one has to store much more data than what would otherwise be necessary, simply to solve the problem of latency (since there is no connection that permits communication in real-time).

Another manner to solve this problem, is to use another transport for HTTP than TCP, for example XMPP. This protocol (standardized by the IETF, like HTTP) gives all participants a global identity, permits authorization of messages and, because of this, also distributed transactions, real-time communication, it bypasses firewalls in a secure manner, and is scalable to many billions of devices (today, already billions of devices are using it).

More information can be found here:

The project could include:

*         Implementation of the HTTP over XMPP protocol, and the HTTPX URI scheme, as plug-ins to browsers (Mozilla, Chromium), web servers (Apache), and perhaps media players (VLC)

*         Implementation of the same into a public SPARQL endpoint (or SPARQL database), or use an existing one. (Clayster platform already supports it)

*         Implementation of an ontology for the Internet of Things.

*         Measurement of efficiency.

*         Simulation of huge networks.

*         Security analysis.

*         Proposal of extensions.

The BIOTIC protocol (Binary Internet Of Things Interoperable Communication) proposes to solve many of the problems that exist today for the internet of things, and it also proposes several new innovative solutions. To learn more, look at the documents published here:

This document describes the reason for creating and using the protocol:
https://dl.dropboxusercontent.com/u/3064761/Biotic%20-%20Executive%20Summary.pdf <https://dl.dropboxusercontent.com/u/3064761/Biotic%20-%20Executive%20Summary.pdf%20>

This document describes the protocol, including new innovative communication patterns:

The project could include:

*         Implementation of BIOTIC-servers, which can be installed in home gateways, on routers or publicly on the Internet.

*         Tiny implementations for microcontrollers

*         Measurement of efficiency.

*         Simulation of huge networks.

*         Security analysis.

*         Proposal of extensions.

If there are any interested parties, I propose that we schedule a meeting to talk about this, and how we can advance to be able to send one or more proposals before the indicated end date.

If you know any other researchers that could be interested, please resend this email to them.

Best regards,
Peter Waher
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/iot/attachments/20141219/23615c53/attachment-0001.html>

More information about the IOT mailing list