INSPIRE: Exploiting Cross-Disciplinary Synergies For Efficient Perishable Commodity and Information Distribution

Temple University, Philadelphia PA

Investigators

Abstract

This INSPIRE award is funded by the Networking Technology and Systems (NeTS) in the Computer and Network Systems Division (CNS) in the Directorate for Computer and Information Science and Engineering (CISE), the Chemical, Bioengineering, Environmental, and Transport Systems (CBET) Division in the Directorate for Engineering (ENG), and the Office of Integrative Activities. The disciplines of commodity distribution logistics and information networking have been studied extensively for many decades but almost exclusively in their own domains. In the logistics area, the distribution of fresh food and other perishable commodities is becoming critical but the perishability has not been well integrated into the logistics. The purpose of this proposal is to expose synergies between Perishable Commodity Distribution Networks (PCDN) and Information Networking (IN) involving transport of delay sensitive content, and exploit them to solve complex problems in both fields. Such a synergistic study is particularly timely because of (a) rapid infusion of information and communications technologies (ICT) in the logistics space, and (b) rapidly increasing carbon footprint in both areas. The Principal Investigators' (PIs') preliminary examination of perishable commodity distribution and computing in general have shown a number of synergies and inspired a number of cross-domain methods that this proposal will examine in detail. In particular, the content centric networking (CCN) paradigm is being actively examined in the context of next generation networks, and inspires similar approaches for the perishable commodity distribution that the proposal will examine. Bundling and unbundling of packages plays a crucial role in logistics and inspires not only the examination of bundling of multiple types of perishable commodities for higher efficiency but also bundling of contents in data center network via optical networks that are particularly suited for bulk data transfer. Virtualization is a well-known technique in cloud computing systems and the project will explore it in the context of PCDN. The much higher complexity of PCDN not only brings new challenges but also inspires more sophisticated virtualization schemes for computing involving latency sensitive processing. Physical constraints in logistics such as human aspects of scheduling truck drivers and varying produce availability make the efficiency of PCDN particularly low, estimated to be in the teens. The project not only examines mechanisms for better accommodation of these constraints but also examines a new data center optical networking scheme inspired by such solutions. Finally, the proposal seeks synergies between environmental impact tradeoffs in PCDNs (e.g., minimizing food spoilage vs. reducing empty truck miles) and data centers (e.g., minimizing cooling vs. maximizing server usage). This research is examining these and other issues and developing techniques for them for both PCDN and IN. The broader impacts of the research are the development of cross-domain techniques that can be used for better resource efficiency in both PCDN and IN areas. These methods have the potential to positively impact the environmental sustainability in both of these areas. For instance, even in developed countries, a huge percentage of fresh food (perhaps approaching 50%) is discarded at some point in the supply chain, which has a huge environmental impact in terms of agriculture (e.g., water, fertilizer, electricity use, fertilizer runoff, etc.), food processing, and distribution of the food. Thus an even small improvement in the management could result in rich environmental dividends. The IN side also has a consistently increasing carbon footprint as ICT is infused into all aspects of human life including physical infrastructures and human health. Thus the efficiencies in the IN domain could also have significant positive impact.