2009 International Conference on Future Networks
Speckled Computing: Evolution and Challenges Sayed Chhattan Shah, Fida Hussain Chandio, and Myong-Soon Park∗ Brunel University, United Kingdom
Internet Computing Research Lab Department of Computer Science and Engineering Korea University, Seoul, Korea {shah, myongsp}@ilab.korea.ac.kr
[email protected]
is involved in digital signal processing aspects of project. St. Andrews is conducting research on power consumption aspects of the project. Their main objective is to design a battery that conserve power and extend lifetime of SpeckNet. Napier University is putting its efforts towards self-organizing systems so that technology can easily be deployed with none or minimal human intervention.
Abstract - Speckled computing is an emerging technology in which data will be sensed in a minute, ultimately around one cubic millimeter, semi conductor grain called speck. Wireless network of thousands of specks called SpeckNet and distributed processing of information on programmable network is termed as Speckled Computing. Specks are not new, but the re-design of sensor nodes at nano scale. These nodes will operate with out any fixed infrastructure and are intended to be deployed in large quantity to increase overall throughput of the system. Furthermore specks can be deployed to places that are difficult to reach. These important features impose various requirements on specks. This paper begins with brief overview of speckled computing and continues with discussion on the evolution of sensor technology along with key research projects. This paper also reviews key challenges related autonomic and ad hoc nature of specks.
Specks are not new. In past couple of projects were started to design and develop sensor nodes at micro-scale. The Speck is the continuation and improvement upon work done on previous projects. The goal of the research is to redesign components of a sensing, processing, and wireless communication at nano scale, allowing the complete package to measure just 1x1x1 (mm). Although specks are very limited in terms of storage, processing power and communication but can be powerful as a part of collective system when connected to each other forming a SpeckNet.
Keywords: Specks, Motes, Mobile Ad Hoc Networks and Autonomic Computing
1. INTRODUCTION
This technology has wide range of applications in various areas ranging from healthcare to defense. Due to extremely small size of sensing nodes, specks, this technology will introduce new class of applications especially in health care. It could be used to repair damaged body cell or to break stones in Kidney, Ureter, and Urinary Bladder and in other organ where stones can be formed or to dissolve blood clots, which can cause brain stroke, heart attack, and pulmonary embolism.
Speckled computing is an emerging technology in which data will be sensed in a minute, ultimately around one cubic millimeter, semi conductor grain called speck [1]. Wireless network of thousands of specks called SpeckNet and distributed processing of information on programmable network is termed as Speckled Computing [2]. A speck has capability to process, sense and communicate in range of few centimeters. They are intended to be autonomous, mobile, and can be proactive in nature.
The aim of this paper is to provide an overview on evolution of speckled computing and to review key challenges related to autonomic and ad hoc nature of specks. It presents key research projects and programs that are based on sensor technology and shows how this technology has evolved over the time.
Speckled computing is currently being researched and developed by a group of five universities where each group is involved in different aspects of this technology. The ultimate goal of each of these is to reduce the size of each and every part of the speck so that the complete package can be designed and developed at nano scale.
The rest of the paper is organized as follows. Section 2 is devoted to evolution of sensor networks. Section 3, 4 discusses the requirements and issues related to autonomic and ad hoc nature of specks. Section 5 focuses on humanspecks interaction while section 6 ends with conclusion.
Edinburgh University is leading the research and is engaged in the design and development of network software for unreliable, densely distributed and highly resource constraint network, SpeckNet. This network software will take care of code partition, migration, parallelization, scheduling, and resources allocation issues. Glasgow University is contributing its effort towards the miniaturisation of antennas for Specknet while Strathclyde 978-0-7695-3567-8/09 $25.00 © 2009 IEEE DOI 10.1109/ICFN.2009.30
∗
181
Corresponding Author
against variety of opponents and threat situations, these systems were intended to gather valuable pieces of information, widely scattered throughout various forces, partners, and allies. In the beginning of 21st century, advances in computation and radio communication made possible to design and develop small and inexpensive sensors based upon Micro Electro Mechanical System (MEMS) technology, wireless networking and inexpensive low power processor. Again Defense Advanced Research Projects Agency, (DARPA) started a research program SensIT (Sensor Information Technology), as a sensor network to leverage the latest technological advances. The goal of the SensIT program was to create software, algorithms, and all key information technologies for distributed micro-sensor networks to be used in battlefield surveillance, reconnaissance, and targeting. This program was based on the concept of cheap, small, interactive and programmable smart devices with embedded processing, and storage capability along with short-range wireless communications. These devices were intended to be quickly and flexibly deployed in ad hoc fashion and in highly dynamic environment for varying missions, potentially in very large numbers, hand emplaced or air dropped, on buildings and bodies, on vehicles, and on ground and under water [6] [10] [11].
2. EVOLUTION OF SENSOR NETWORKS Historically, military has played the leading role in research and development of sensor networks. Development of sensor networks dates back to 1930 when scientists at U.S Naval Research Laboratory in Washington, D.C., used radar to detect aircrafts. During cold war, U.S installed Distant Early Warning (DEW) network of long-range radar to warn against bomber attacks and UK installed the Ballistic Missile Early Warning System (BMEWS) to warn against intercontinental ballistic missiles (ICBMs) [17]. In 1983 President of United States proposed a research program, Strategic Defense Initiative (SDI), for developing an Anti-Ballistic Missile (ABM) defense system. The program was intended to provide total protection against nuclear attack by intercepting incoming missiles in midcourse. Although this program was abandoned due to budgetary problems but research and development in this program created significant technological advances in computer systems, component miniaturization, sensors and missile systems that form the basis for current systems [12] [18]. Over the time, systems were evolved and several new research projects were started to cope up with threats and to maintain minimum deterrence. Technology got commercial acceptance and was actively used in air traffic control, satellites, space exploration, remote sensing, submarine tracking and detection, automobiles and in many other military and civilian applications.
In late 1990s, United States and its allies launched a project named Future Solider to equip ground-based combat soldier with an integrated set of high-technology uniforms and equipments. These equipments were intended to be linked to an array of real-time and archived battlefield information resources. The Future Soldier equipment was tested in the 2003 invasion of Iraq and currently being assessed for possible deployment to Iraq [14].
The modern research on sensor networks was started around 1980s with “Distributed Sensor Network” program at Defense Advanced Research Projects Agency, DARPA. The objective of the program was to develop and extend target surveillance and tracking technology in systems that employ multiple distributed sensors and processing resources. The system was intended to gather and process information from spatially distributed sensing nodes in order to provide a complete and accurate picture of surveillance area. The major technology components of DSN were identified as autonomic sensing nodes, data bases, and processors distributed throughout an area and interconnected through communication system. Since very few technology components were available off the shelf, researchers at Carnegie Mellon University (CMU), Pittsburgh, PA, Massachusetts Institute of Technology (MIT), Cambridge and Advanced Decision Systems (ADS) developed network software and algorithms for data acquisition, acoustic tracking, and signal processing for distributed sensor network [6]. . Although earlier research was not focused towards the small sensor networks because technology was not quite ready but with success of DSN program, the military planners recognized the importance of small sensor networks especially in network-centric warfare [13] [19]. These systems were intended to provide operational commanders the situational awareness - information about the location and status of enemy and friendly forces to prosecute their missions more effectively. To combat
In 1997 University of California, Berkeley started a project named “Smart Dust” or “Mote”. The same project was funded by Advanced Research Projects Agency, (DARPA) in 1998. Motes are tiny, self-contained, batterypowered computers with radio links, which enable them to communicate and exchange data with one another, and to self-organize into ad hoc networks. Various firms are producing motes that are available with processing power of few kilohertz to few megahertz and with communication range of several hundred feet. They can transmit approximately 40Kbps and are measured in cubic centimeters. Researchers are working on several projects to reduce the size of motes while increasing their processing power and battery life. Several operating systems for motes have been developed and are available commercially. Future Force Warrior part of Future Combat System project is other project that seeks to create a fully integrated individual combat system with enhanced human performance. The integrated individual combat system will comprise weapon system, protection suite, navigation, and information processing and communication systems. The individual combat system will be interoperable with C4ISR [21] system for communication and collaboration with command and control center and for collection and dissemination of information about battlefield. The first phase of the project involves development of technologies 182
to help reduce the soldier's fighting load and power requirements and improving soldier’s protection, lethality, and environmental and situational awareness. [9] [20] [22] MIT Institute for Solider Nano-Technologies also started a project named “Solider of the Future” to develop individual combat system based on nanotechnology enabled materials and devices. The main objective of program is to design and develop system with capability of threat detection, threat neutralization, concealment, realtime automated medical treatment and protection against bullets and chemical and biological agents. The system is intended to enhance solider performance and reduce weight of equipment [23] [24]. Recent revolutionary advances in integration, miniaturization, sensor, computing and communication technologies have caused a significant shift in sensor network research. Currently, research is being held in various dimensions. One part of research is devoted towards design and development of nodes at micro and nano scale. The other part of research is devoted towards making these devices autonomous and proactive in nature, enabling the transitions from today’s interactive systems to proactive environment that anticipate our needs and act on our behalf.
3. AUTONOMIC SPECKS Specks are extremely limited in terms of processing, storage, communication and battery power. They are usually deployed in large quantity to increase overall throughput of the system. This creates complexity in overall network, which is hard to control manually. Manual control is also time-consuming, expensive and error prone. Specks are also intended to operate under varying conditions and may interact with diverse systems to exchange information. These features impose following requirements on specks:
Specks are intended to operate in hostile environment, may interact with other services or may operate in smart spaces, so specks should be expert in self-protection.
Specks interact with each other and with external environment to exchange information and to perform tasks so they should be capable of establishing and maintaining relationship with other systems.
Specks have very limited capabilities so they perform task by collaborating with each other. This collaboration incurs communication, which is very costly and also consumes more power than computation. This requires that specks should reduce the cost of communication by avoiding unnecessary exchange of data and should manage their resources in efficient manner.
Specks are nano scale devices that can be deployed in places that are extremely difficult to reach which require that they should be able to organize themselves.
Designing an effective system that could meet above requirements is the most challenging issue. The emerging paradigm, autonomic computing offers a promising solution to above requirements with none or minimal human intervention. The autonomic computing is inspired from the biological systems that regulate and protect their bodies and are able to cope with environmental changes, external attacks and internal failures. The autonomic computing is active area of research. Many industries and academia are conducting research on various aspects of this technology. IBM is leading the research in this field and has defined following four aspects of autonomic systems [15]:
Commercial companies such as Ember, Crossbow, Sensoria and SmartTech are now building and deploying small sensor nodes and systems for military and civilian applications.
Self-Configuration: Automatic configuration components in accordance with high-level policies.
of
Self-Healing: Automatic discovery, and correction of faults
They should be able to configure themselves under varying conditions.
Self-Optimization: Automatic monitoring and control of resources to ensure the optimal functioning with respect to the defined requirements
They should be adaptable according to environment. Adoption can cause them to shift between drastically different modes of operation.
Self-Protection: Proactive identification and protection from arbitrary attacks.
Specks operate based on their batteries that have extremely limited life. Once the batteries are consumed these devices cannot function anymore. Therefore they should continuously monitor themselves to optimize their power consumption in order to maximize total network lifetime.
As discussed above, specks are extremely limited in environment so the main issue is to decide the degree of autonomy in specks. There is a tradeoff between degree of autonomy and processing, the higher overall system’s autonomy, the less human involvement and greater the need for extensive and sophisticated processing inside the system. Designing a fully autonomic system is a challenging issue especially in such extremely limited environment. Furthermore designing individual autonomic specks and composing system out of these specks in such a way that the system as a whole should be autonomic is other challenge that requires further attention. There are
Specks also should be able to heal, to recover from failure. This requires that they should monitor themselves for the problems and should discover the solutions to those problems.
183
various research projects under development but they are targeted towards large-scale system.
6. CONCLUSION Speckled computing is a generic technology for ubiquitous and pervasive computing that will revolutionize the way we live and interact with each other and with physical world. This technology has wide range of applications in various areas ranging from healthcare to defense. Due to extremely small size of sensing nodes, specks, this technology will introduce new class of applications especially in health care. It could be used to repair damaged body cell or to break stones in Kidney, Ureter, and Urinary Bladder and in other organ where stones can be formed or to dissolve blood clots, which can cause brain stroke, heart attack, and pulmonary embolism. In this paper an overview of speckled computing and an evolution of this technology were discussed. Various key research projects were reviewed to understand the evolution of the technology and related work. Finally key requirements and issues were presented with respect to ad hoc and autonomic nature of specks.
4. MOBILE AD HOC SPECKNET SpeckNet do not rely on any fixed infrastructure instead specks rely on each other to keep the network connected. There is no central authority for administration and management. Every speck is responsible for keeping the identity and location of its neighbors to support processing and collaboration. This infrastructureless environment introduces following issues. 4.1 Dynamic Topology In SpeckNet, specks are mobile and can be in motion in the region without any restriction. Due to this movement the coverage area is also going to change. Furthermore, new specks can join the network any time and the old ones can leave at any time whenever they will. Due to these factors the topology of the SpeckNet changes very frequently. This issue has a major impact on the way data travels in a network. Dynamic routing implies that changes in the topology may also lead to changes in the routing paths of data. The network of specks should be able to cope with dynamic nature of topology.
7. REFERENCES [1] D.K Arvind, K.J Wong “Speckled Computing – A Disruptive Technology for Network Information Appliances” IEEE International Symposium on Consumer Electronics, 2004
4.2 Limited Physical Security [2] D. K. Arvind “Speckled Computing”, Invited Paper, Proc. Nano 2005, Vol. 3, pp 351-54, ISBN: 0-9767985-2-2 May 2005
SpeckNets are prone to malicious attacks due to lack of centralized authority, there is no one in the network who takes care of the security. Lack of any centralized authority makes these dynamically changing wireless structures very vulnerable to penetration, eavesdropping, spoofing, denial of service attacks, interference, etc. Security is a major obstruction in the applications running on SpeckNet. The available algorithms for encryption are also not suited for the specks because of their extremely limited computational power.
[3] Kasim Rehman, Frank Stajano, George Coulouris, “Interfacing with the Invisible Computer” NordiCHI 10/02 Århus, Denmark © 2002 ACM ISBN 1-1-58113-616 [4] K.J Wong, D.K Arvind, N Sharwood –Smith, A Smith “Speck-based Responsive Environments” ROCEEDINGS OF THE IEEE 05, 0-7803-8920-4 [5] Mathew Leach, David Benyon “Interacting with Speckled World” ADPUC'06, Melbourne, Australia. ACM 1-59593-422-7
4.3 Privacy Issues Probably, the most important problem is that speck-based computing environments pose serious privacy risks. By watching everything a user does, these systems have the potential to leak all our information and actions to others unknown to us. Designing the solution to deal with privacy issues is a big challenge for these extremely limited nodes.
[6] Chee-Yee Chong, Srikant Kumar “Sensor Networks: Evolution, Opportunities, and Challenges” ROCEEDINGS OF THE IEEE, VOL. 91, NO. 8, AUGUST 2003. [7] Speckled Computing, http://www.specknet.org/about, accessed on March 2008
5. HUMAN SPECKS INTERACTION [9] Future Force Warrior, http://ffw.natick.army.mil/index.htm, accessed on April 2008
Specks can be invisibly integrated into our environment to monitor surroundings or can be embedded in our bodies to perform diverse tasks. People may wish to interact with the specks to re-assign a task or to replace a speck depending on type of application. Interaction with specks may involve a person interacting thousands of specks concurrently. Design of the interface for interaction between human and specks is a big challenge for applications that involve human intervention [3].
[10]DARPA’s SensIT program, http://www.sainc.com/sensit/goals.htm, accessed on March 2008 [11]Sensor IT program, http://www.fbodaily.com/cbd/archive/1999/12(December)/ 15-Dec-1999/Asol009.htm, accessed on March 2008
184
[12] US Strategic Defense Initiative, http://www.fas.org/spp/starwars/offdocs/m8310017.htm, accessed on March 2008 [13] David S. Alberts, Richard E. Hayes “Power to the Edge: Command and Control in the Information Age” [14] Future Solider, http://en.wikipedia.org/wiki/Future_Soldier, accessed on April 2008 [15] Jeffrey O. Kephart, David M. Chess, IBM Thomas J. Research Center “The Vision of Autonomic Computing” 2005 [16] Roy Want, Trevor Pering and David Tennenhouse Intel Research, “Comparing Autonomic & Proactive Computing” IBM Systems Journal Q3 2002 [17] Blastic Missile Early Warning System, www.fas.org/spp/military/program/track/bmews.html, accessed on April 2008 [18] “Strategic Defense Initiative”, Microsoft ® Encarta ® Reference Library [19] David S. Alberts, John J. Garstka, Frederick P. Stein “NETWORK CENTRIC WARFARE: Developing and Leveraging” Information Superiority [20] Future Force Warrior, http://www.defenseupdate.com/features/du-4-04/FFW.htm, accessed on April 2008 [21] “C4ISR FOR NETWORK-ORIENTED DEFENSE” © Ericsson Microwave Systems AB 2006 [22] Phil Copeland, “Future Warrior Exhibits Super Powers” American Forces Press Service
[23] Institute for Soldier Nanotechnologies, http://web.mit.edu/ISN/aboutisn/index.html, accessed on April 2008 [24] Army Collaborate on Military Nanotech, http://www.washingtontechnology.com/online/1_1/208431.html, accessed on April 2008
185
