Sunday, June 22, 2014

Delay Tolerant Networks - When Transportation Trumps the Internet

It has been fabulous to be back in Sweden where I am continuing my third year appointment as a Visiting Professor at the School of Business, Economics and Law at the University of Gothenburg. I arrived in Europe on May 11. 

Being based in Europe for a couple of glorious weeks allows for some great conference chaining which I have been doing and have also been highlighting on this blog in several previous posts.

There has also been a North South theme to this visit since I have given talks at workshops and conferences, since my arrival, in Erice, Sicily (south), came back to Sweden (north), then ventured to Taormina, Sicily and  Catania, also in Sicily (south) and onwards to Chalkidiki,  Greece (south). Back in the north now in Gothenburg.

As a researcher and a very frequent flier, who also works on networks and their numerous and fascinating applications, including transportation and communications, practical, real-world  experiences reinforce my research and vice versa! Plus, you acquire great stories to share with your students in classes.

Before flying to Europe, we were finishing up a paper on Delay Tolerant Networks, and, in the meantime, we have also completed another one. This is a collaboration initiated by a doctoral student in engineering, who is from Colombia, and who is visiting my Co-PI on our National Science Foundation project, Network Innovation Through Choice, Professor Tilman Wolf of the College of Engineering at UMass Amherst. The doctoral student is Luis Marentes and he sat in on my SOM 825 course this past semester on Variational Inequalities, Networks, and Game Theory.

In the beautiful parts of southern Europe - I can't resist sharing some photos of Sicily and Greece where I spoke at below, there, nevertheless, were major issues with the Internet!  Even when it was up it might take as long as an hour to access an email message and, as for tweeting 140 characters, sometimes we waited for 24 hours, and the tweet was still not transferred.  But we were patient and the beauty and the caliber of the  conferences and people made up for the technology shortcomings. I will share a story at the end of this post.

Our work on Delay Tolerant Networks, is inspired, in part, by that of Dr. Alex "Sandy" Pentland of MIT, whom we have hosted at the Isenberg School of Management at UMass in our INFORMS Speaker Series, and whom I have interacted with on other occasions. I am a huge fan of Sandy's and was thrilled when he was ecleted to the National Academy of Engineering this year.

Pentland's paper, DakNet: Rethinking Connectivity in Developing Nations, published in Computer in 2004, speaks of connecting remote villages in India and Cambodia through a combination of wireless and asynchronous means.  

When I tell people that there are parts of the globe in which residents desperately need information that we take for granted and that which we access through the Internet but they lack digital connectivity because of market and pricing forces (including poverty) and can benefit from using transportation modes (from busses to mules - I am not kidding) to secure the information, I receive expressions of shock!

In our first paper, "Overcoming Economic Challenges of Internet Operators in Low Income Regions Through a Delay Tolerant Architecture with Mechanic Backhauls," Luis Marentes, Tilman Wolf, Anna Nagurney, Yezid Donosoa, and Harold Castro,  we study the following. Regions with low population densities and low income, which may be rural, are far behind in Internet access as compared to their urban counterparts. Economic unsustainable deployments have been suggested as one of the factors with high negative impact. Researchers using the Delay Tolerant Network Architecture (DTN) have established a less expensive alternative, but they do not provide guidelines on how to price the new possible services nor any indication as to operators’ profitability. These two points are the contributions of this paper. Based on a continuous pricing model, we show that projects using the DTN architecture have a positive net present value using real demand data.

In this paper, we also explore how far the use of this technology is going to help in attaining profitable deployments with real and delay tolerant services. Our findings indicate that although it is not possible to support the investment reported in a previous actual deployment, a great portion of the investment is actually covered by the predicted gross profit of our model. We hypothesize that the remaining value is affordable if service differentiation is enabled by means of a new architecture. Three components to test the hypothesis are introduced: (1) the pricing elements to form service level agreements, (2) a general overview of the architecture requirements, and (3) a novel pricing model. In this regard, our results show a positive effect of service differentiation which improves profits by 10% without any customer additional payment.

In our second paper, "Towards Pricing Mechanisms for Delay Tolerant Services," Luis Marentes, Tilman Wolf, Anna Nagurney, and Yezid Donoso, we again note that one of the applications of the Delay Tolerant Architecture (DTN) is rural networks. For this application researchers have argued benefits on lowering costs and overcoming challenging conditions under which, for instance, protocols such as TCP/IP cannot work because their underlying requisites are not satisfied. New responses are required in order to understand the true adoption opportunities of this technology. Constraints in service level agreements and viable alternative pricing schemes are some of the new issues that arise as a consequence of the particular operation mode. 

In this paper, we propose a novel model for pricing delay tolerant services, which adjusts prices to demand variability subject to constraints imposed by the DTN operation. With this model we also show how important parameters such as channel rental costs, cycle times of providers, and market sensitivities affect business opportunities of operators.

In both of these papers, the consumers respond to the average time of the delivery of the services, which could consist of a combination of communication and transport modes!

As for our journeys in southern Europe, we ended up, in Chalkidiki, Greece taking the ferry to the neighboring town of Marmara (transport) and back to seek a decent WiFi connection.

And we were successful. Decent WiFi was found at a restaurant at which we also had a fabulous meal - it was so good that I blew kisses to the cooks and then they surprised us with a tray of fresh cherries and exotic, delicious Greek desserts!

We then, the next day, explored another mode of transport to access the Internet - we walked rather than taking the ferry to Marmara and it took us about the same amount of time - plus the beach views were glorious. And for those news junkies there, we even found a small shop where we purchased the same day edition of The International New York Times, which was printed in Greece. The owner is British.