The MMHC Project

The purpose of the MMHC Project is to provide a real-world support to collaborative networking, based on the dynamic and autonomous management of networking capabilities; in particular we aim to support
• infrastructure/adhoc hybrid connectivity,
• multi-hop paths,
• multi-path selection,
• heterogeneous interfaces.

Let us stress our purpose is twofold:
1. providing an effective middleware implementation supporting collaborative scenarios
2. providing an unifying model able to consider infrastructure/adhoc connectivity via single/multi hop paths and based on heterogeneous wireless interfaces.

While these pages briefly sketches our proposed model and its implementation, we provide even a deep investigation of the current literature based on the proposed model. If you are interested in our model and survey, please refer to the paper below.

Paolo Bellavista, Antonio Corradi, Carlo Giannelli: A Unifying Perspective on Context-aware Evaluation and Management of Heterogeneous Wireless Connectivity, IEEE Communications Surveys and Tutorials, Vol. 13, No. 3, pp. 337-357, Third issue 2011.
[pdf paper @ IEEE]

While, in principle, the wireless interface technologies to implement MMHC environments are already there (heterogeneous non-interfering wireless cards, simultaneous usage in both infrastructure and ad-hoc modes, …), the most common service scenario today is much more simplistic than what envisioned above. It is based on single-hop connections between clients and infrastructure-based APs/BSs, explicitly chosen by users once for all their applications and independently on mobility-aware considerations.

To favor the realization and wide acceptance of MMHC, we claim the crucial need of novel, effective, and mobility-aware middleware supports. We claim that MMHC middleware should operate lightweight management decisions to quantitatively evaluate the locally available single-hop paths; based on this first local evaluation, to the purpose of limited overhead and high scalability, the middleware should consider only the limited set of deriving multi-hop opportunities; once identified the few appropriate multi-hop paths to activate, the middleware should operate the consequent configuration/routing decisions in a decentralized and self-organized way. To achieve that goal, our practical experience on heterogeneous wireless systems suggests exploiting a few, simple, and mobility-aware indicators to effectively give a coarse-grained estimation of i) MMHC reliability (expected durability) and ii) MMHC quality (expected throughput). By considering these concise indicators and by carefully taking into account the performance/optimality tradeoff achievable via local/global management decisions, it is possible to enable effective management middleware with feasible performance for most application domains. Finally, we claim the suitability of a middleware-layer approach, with no impact on underlying protocols and interface implementations, thus facilitating rapid deployment in already available wireless environments.

By following these guidelines, we have designed and implemented a novel MMHC middleware that, based on coarse-grained durability/throughput estimations, fully exploits the capabilities of commercially available interfaces. Our middleware bases its management decisions on two main evaluation procedures. On the one hand, a local procedure pre-selects a subset of the available single-hop paths by considering both APs/BSs and mobile peers; its primary purpose is to identify potentially durable single-hop connections by exploiting only simple monitoring data locally available at client interfaces. On the other hand, a global procedure determines the most suitable MMHC paths, by processing lightweight distributed management data and consequently performing configuration/routing on collaborating nodes. The reported experimental results demonstrate the feasibility of our mobility-aware middleware approach and the effectiveness of our simplifying assumptions on durability/throughput estimators. In addition, we hope that the availability of our middleware prototype as an open-source tool for the community could be useful to accelerate and further stimulate the practical research work in the field.


Privacy Policy