NorVAL : a methodology and a system to evaluate the "openness" of a cluster of firms

This paper presents a methodology and a set of tools that we are developing to define and assess the level of “openness” (http://en.wikipedia.org/wiki/Openness) of a cluster of firms. As the complexity of business processes grows, modern firms tend to organize themselves in networks of task oriented delocalized units. The goal of managing complexity drives this trend, but a truly modular organizational setting - where units interact only through predefined transactions - enables parallel work and may improve the flexibility with respect to the uncertainty of the future [Baldwin 2003]. Artifacts (atoms) have to be physically transported (using some kind of vehicle) while information (bits) can obviously be digitally transmitted through information networks (e.g. telephone, Internet, etc.). From now on we will focus on the digital aspect. Intuitively, the nature (format and/or packaging) of what is exchanged has an impact on the value of the information exchanged. In fact, given the same content, standard and open formats bring more value than proprietary and closed ones. One motive for this is that information value is directly proportional to its accessibility/availability, and open formats are more accessible than proprietary ones. Our methodology, codenamed NORVAL (NetwORk eVALuation), analyzes many pieces of information (e.g. files and packets) scattered through the network of a group of interconnected (usually via Internet+VPNs) firms. After the analysis we compute an “openness value” based on a weighted average of all the information gathered. The first stage of NORVAL is represented by the definition of an abstract (reduced from the real world) model of a network of firms. We describe a network of firms as a cluster of interlinked networks. Each network is described as a cluster of nodes connected by links through which packets flow. Packets form protocols transporting information in some format. The idea is simple to describe: 1) define values (weights) for every class/subclass in the abstract model (example below); 2) gather information about all the data stored and flowing in the network of firms; 3) compute value. For example, speaking about documents (files), we may arbitrarily define a range of values for different types of document formats, e.g.: ODT (0.8), DOC (0.5), OTHER (0.3). Then we gather information about the number of documents in the network, e.g.: ODT (1230), DOC (3500), OTHER (45). So that the value of the TextDocument class in the example network would be: 0.8*1230+0.5*3500+0.3*45 = 2747.5. Thus, for example, by only converting every DOC instance to an ODT instance me may reach a much higher value of 3797.5. Of course the mathematical formula can be changed (e.g. normalized) but the general idea remains the same. While the idea is straightforward, its implementation may be a problem... We are implementing it by combining “nmap”ping, scanning and parsing techniques (using FLOSS tools, glued together by scripting) to gather information about: files stored (reached through NFS/CIFS/FTP/etc.), services running (listed through nmap or similar scanning software). At the moment we don't do sniffing but it may be useful in the future. We are also devising a parsing technique to assess the cost (in terms of feature loss) of converting any MsOffice document into an OpenOffice one. An initial description of our work can be found in [Trentini 2007]. (The authors gratefully acknowledge financial contributions from the FIRB project “International fragmentation of Italian firms. New organizational models and the role of information technologies”, a research project funded by the Italian Ministry of Education, University and Research).