A store-carry-process-and-forward paradigm for intelligent sensor grids
Introduction
Grid computing [33], [15] has consolidated as a technology capable of solving some of the most challenging scientific projects of our century. The needs of these projects usually include complex computation of data obtained from different sources and stored in large storage resources. The main goal of grid computing, precisely, is to share these resources among different institutes and virtual organizations across high-speed networks and distribute and coordinate its processing.
Wireless sensor networks (WSNs), on the other hand, is a technology that can be very useful when it comes to acquiring and transporting data collected in widely spaced areas. These networks consist of different nodes carrying different sensors along with autonomous computational devices which transmit data through the network to some specific locations or data sinks.
This article analyzes how both technologies, grid computing and wireless sensor networks, can be combined into an integrated WSNs and computer grid infrastructure allowing new functionalities. The corner stone of this conjugation is using delay and disruption tolerant networking (DTN) concepts [14] along with mobile code to create an intelligent grid network capable of routing and managing processes depending on the context. Some other recent proposals, which will be further described in Section 2, integrate WSNs and grid computing, as well. However, our proposal comes from the network perspective. We consider WSNs nodes as intermittent connected nodes, containing asymmetric bandwidths, long and variable latencies and ambiguous mobility patterns. This new perspective contributes to the creation of a novel concept of intelligent grid computing networks, going beyond the possibilities of the reviewed literature in some current scenarios, and providing promising prospects for supporting future grid services.
The routing decision making and execution policies travel with the messages, instead of being static and exactly the same for all nodes. These policies, in the shape of mobile code, can take into account the context of the nodes to choose the behavior that fits best in each situation. All in all, the system acts more like an ant colony, with differentiated autonomous parts acting locally but with a cooperative aim, rather than a traditional and more inflexible system. Thus, using mobile code makes the grid network an intelligent system, pliable enough to adapt to new scenarios of grid computing. However, the proposed system cannot be considered a silver bullet for all grid computing; in highly connected grids, with low latencies and where data does not need to be processed before getting to the execution destination, mobile code would introduce an unnecessary extra overhead and other unwanted side effects.
The original contributions of this paper are: a grid computing model based on mobile code to allow intermittently connected wireless sensor networks to seamlessly coexist with other traditional connected services and a grid computer service which transparently gives access to a general purpose multi-application mobile robot node sensor network.
This paper is organized as follows: Section 2 is a state of the art of the combined use of grid computing and wireless sensor networks, having a closer look at those ones which suffer from intermittent connections and those using mobile code. Section 3 presents our proposal, an intelligent system following the store-carry-process-and-forward paradigm. In Section 4 we present a delay tolerant architecture for grid services which are transiently unavailable. In Section 5 we analyze how to include intermittent connected networks in traditional grid computing systems using mobile code. Services such as computing, storage, information service and monitoring will be discussed in detail. Finally, Section 6 presents the conclusions we have come to.
Section snippets
Background
There are several efforts already published on the integration of wireless sensor networks and grid computing. Studies like [30] and [23] propose different ways of extending the computing grid paradigm to allow the integration of wireless sensor networks and grid computing infrastructures. This section analyzes the state of the art of other technologies: mobile code and DTN protocols which we believe can extend and improve this integration. In Fig. 1, research overlapping of the four
Integrating delay/disruption tolerant WSNs in grid computer infrastructures using mobile code
In this section we propose a way to integrate intermittently connected WSNs in computer grid infrastructures. On one hand, the data acquired by the WSNs can be processed and stored using traditional grid services inside connected regions. On the other hand, and a more challenging option, sensors can be considered as grid sources of data commanded by the very grid users themselves. We propose the job behavior that evolves beyond the traditional read-process-and-storage model – that is reading
Implementation
The execution environments on which intelligent messages run are called platforms. We need to implement on every sensor grid node an execution environment platform to let intelligent messages carry grid level information to run on. In order to implement intelligent messages described in the previous sections, we need some special platforms: besides being capable of executing code, platforms must allow the code to be forwarded from one sensor node to another, stopping their code, resuming its
A practical example: a general purpose grid multi-application robot sensor network
In this section we present as an example of our proposal, a way to include a general purpose multi-application mobile node sensor network in an existing computer grid infrastructure.
We have implemented a computer element job manager which transparently gives grid access to a robot wireless sensor network. This job manager creates an intelligent message for every grid job submitted. Users using gLite’s job description language [20] choose among three different types of intelligent message
Conclusions and future work
We have presented in this paper an intelligent system to transparently integrate intermittently connected wireless sensor networks to computing grid infrastructures. We have presented a new paradigm called store-carry-process-and-forward, which proposes how to process data while the information is stored, waiting for the DTN information to be forwarded.
From another hand, as a practical example of our proposal, we have introduced a computer element service which provides access to a general
Acknowledgements
Carlos Borrego and Sergi Robles gratefully acknowledge the support from Ministerio de Ciencia e Innovación, Spain. This work was supported in part by Grants TIN2010-15764, FPA2007-66152-C02-00, TIN2010-15764, FPA2010-21919-C03-02 and FPA2007-66708-C03-02 from Ministerio de Ciencia e Innovación, Spain.
References (36)
- et al.
Communication in disconnected ad hoc networks using message relay
Journal of Parallel and Distributed Computing
(2003) - et al.
Swarm intelligence based routing protocol for wireless sensor networks: survey and future directions
Information Sciences
(2011) A biological atlas of functional maps
Cell
(2001)The ATLAS experiment at the CERN large hadron collider
JINST
(2008)- H.M. Almasaeid, A.E. Kamal, Data delivery in fragmented wireless sensor networks using mobile agents, in: Proceedings...
- N. Asokan, K. Kostiainen, P. Ginzboorg, J. Ott, C. Luo, Applicability of identity-based cryptography for...
- F. Bellifemine, A. Poggi, G. Rimassa, JADE. A FIPA compliant agent framework, in: Proceedings of PAAM, 1999, pp....
- C. Borrego, S. Robles, Relative information in grid information service and grid monitoring using mobile agents, in:...
- et al.
Large hadron collider tests of the little Higgs model
Journal of Physical Review Letters
(2003) - A. Carzaniga, G.P. Picco, G. Vigna, Is code still moving around? Looking back at a decade of code mobility, in:...
Mobile-C: a mobile agent platform for mobile C-C++ agents software
Practice & Experience Archive
Mobile agent computing paradigm for building a flexible structural health monitoring
Sensor Network Computer-Aided Civil and Infrastructure Engineering
Relay node placement in wireless sensor networks
IEEE Transactions on Computers
Delay-and Disruption-Tolerant Networking
PIVOT: an adaptive information discovery framework for computational grids
Information Sciences
Cited by (8)
Hey, influencer! Message delivery to social central nodes in social opportunistic networks
2019, Computer CommunicationsCitation Excerpt :Softwarecast uses Active-DTN [33], an OppNet solution that consists in extending the messages being communicated by incorporating software code for forwarding, delivery, lifetime control and prioritisation purposes. Active-DTN has been applied in different DTN scenarios such as opportunistic computing infrastructures [34], multi-application mobile node sensor networks [35] and disconnected emergency scenarios [33]. We base our proposal on the study OppNet Profile-cast scheme [36].
Softwarecast: A code-based delivery Manycast scheme in heterogeneous and Opportunistic Ad Hoc Networks
2017, Ad Hoc NetworksCitation Excerpt :More recently, there are interesting studies using mobile code in OppNet scenarios. Borrego et al. [5] propose a new paradigm called store-carry-process-and-forward that uses mobile code to improve the integration of wireless sensor networks and grid computing infrastructures. Additionally, Borrego et al. [3] present a general purpose, multi-application robot-sensor network based on mobile agents that run mobile code on Mobile-C platforms [26].
Buffer management for streaming media transmission in hierarchical data of opportunistic networks
2016, NeurocomputingCitation Excerpt :Moreover, given that the nodes have limited battery power, are sparely distributed, and move frequently, preventing the network from intermittent connection is a fundamental concern among researchers [14]. Generally, the store-carry-and-forward protocol is adopted as the routing mode of the nodes, and this makes full use of the mobility to transmit messages [6]. This network structure is widely used for wildlife tracking, disaster emergency communication, sea or mountain salvage efforts, and new media transfers.
A mobile code bundle extension for application-defined routing in delay and disruption tolerant networking
2015, Computer NetworksCitation Excerpt :There are few studies using mobile code in DTN scenarios. Borrego and Robles [10] propose a new paradigm called store-carry-process-and-forward that uses mobile code to improve the integration of wireless sensor networks and grid computing infrastructures. It proposes the implementation of a delay tolerant grid service, the computer element, to give computing access to an intermittently connected wireless sensor network.
PHDA: A priority based health data aggregation with privacy preservation for cloud assisted WBANs
2014, Information SciencesCitation Excerpt :In addition, some research efforts are paid to investigate data forwarding in health-care systems. Borrego et al. [7] investigate a new paradigm, called store-carry-process-and-forward, based on mobile code to improve the integration of wireless sensor networks and grid computing infrastructures. Liang et al. [12] propose a privacy-preserving emergency call scheme, named PEC, for mobile health-care social networks.