Open Issues of Nomadic Pervasive Computing
Hen-I Yang, Shinyoung Lim, Jeffrey King and Sumi HelalAuthors
Mobile and Pervasive Computing Laboratory
Computer & Information Science and Engineering Department
College of Engineering, University of Florida,
PO Box 116120, Gainesville FL 32611, USA
{hyang, lim, jck, helal authors} @cise.ufl.edu
www.icta.ufl.edu
Abstract. As a user Rroamsing between among islands of smart spaces, additional support within the systems are needed to provide the user with similar experiences and services wherever he or she goes. creates additional challenges on top of the high-complexity of context-awareness and ubiquitous computing. We refer to such environments as nomadic pervasive computing. In this paper, we identify the open issues and suggest necessary system supports under such setting, which we called nomadic pervasive computingto enable such environments. Several events must occur Wwhen users move from one location to another., new bindings The space have tomust be created betweentake the user’s’ preferences and roaming roaming services, and bind them to locally available physical resources. Measures should be taken to guarantee the delivery of emergency services, while provide measures forensuring thesome quality of service for others. The systems have tomust be scalable and secure, while and protecting the users’ privacy where possible. In this paper, wWe provide a summary of the extensive survey of current and previous approaches in related issues, and identify open issues on system support for the nomadic pervasive computingintend for the issues raised here to foster further thought about system support for mobility among smart spaces.
1. Introduction
We believe that in the near future, smart spaces, When people roam around space where ubiquitous computing services are available provides, will become much more widespread. However, it is unlikely that these islands of smart spaces will be joined seamlessly or continuously with each other. Therefore, as users move about in the world, they will have to roam from one autonomous smart space to another. Ideally, users should not have to abandon all the features they have come to expect simply because they move to a new space. While investigating this issue we coined the term nomadic pervasive computing. Nomadic pervasive computing encompasses the federation of isolated smart spaces and the mechanisms that allow users to be continuously served as they move around.
As people get used to new technology, they come to expect its availability, and dependency on the technology grows. This is especially true of smart spaces because their goal is to embed technology into living environments, which can greatly impact many daily activity routines. Consider an elder person with early-stage Alzheimer’s, whose independent living depends on a cognitive assistance service that cues her to finish tasks she has started. When she leaves her apartment to go to the grocery store, she still counts on receiving the same kind of help. This is why we must have nomadic pervasive computing, so that critical services are not lost just because the user leaves her home space.
they would like to have as same as assistance comparing to when they are staying at their own smart space. People will be dismay when they should setup the entire service environment just as the one they used to have before moving to other location. For instance, when foreigners come to US for studying their subjects, they will have series of experience to adjust their lives fit to those in US. For instance, they have to speak in English, learn how to use social facilities and public transportation, and learn how to do daily life in US. It is totally different environment of them. If guidance helps them to quickly and easily settle down in US, it will be very helpful. Just like the guidance, we need help from ubiquitous computing to setup the entire service environment on behalf of people, which means the user will have service environment consistently and seamlessly. It seems ideal from the user’s point of view but it gives ubiquitous system heavy burdens to its resources. The issues deal with this kind of user needs and system requirements of ubiquitous service and computing must be key issues of system support of ubiquitous system. In the age of ubiquitous computing service, the user space is quite variable in regard to users’ devices, terminals, and service providing components. Considering these factors in adjusting the user’s service environment, it is very difficult to be consistent of the user space to the user’s prior service environment. This paper discusses the system support challenges posed by nomadic pervasive computing. These challenges are divided into four categories: binding, quality of service, scalability, and security and privacy. Before we delve into these challenges, a brief survey of research In this paper, we summarize the extensive survey of current and previous system supportrelated to the issues is also provided. issues of ubiquitous computing; make extensive comparison of current and previous research to solve the issue; and raise issues of binding, quality of service, scalability, security and privacy for system support of ubiquitous computing.
2. Related work
One of the prominent researches on ubiquitous computing is the Internet Suspend/Resume (ISR) is a mechanism that layers applies virtual machine technology on top of a location-transparent distributed file system. It belongs relates to the binding issue, that is, the a rapid personalization and depersonalization of anonymous hardware approach forduring transient use of anonymous hardware [47, 58].
ISR assumes transitions occur between workstations. It does not require users to carry portable storage or mobile devices, but the distributed file system requires more functional components for identifying the user, device and service environment [5]. Other systems, such as Aura [14], make use of distributed file systems to expand the capabilities of mobile devices If people carry their own mobile devices in this distributed file system, it provides an alternative way to solve issues on ubiquitous computing system [4]. . However, if people does not carry device in their smart space, the distributed file system requires more functional components for identifying the user, device and service environment [5].
There are different approaches for solving the binding issue. They are for the programming pervasive environment based on the context-driven and service-driven models [1, 2, 3, 11]. Both of the models have shown directions to solve the binding issue. However, we need a bridge to interact with both models, which requires additional functionality in solving the binding issue [10].
There is a different approach to solve the quality of service issue by providing QoS-aware middleware to the system [36]A. The QoS-aware middleware, such as UbiData [3], manages the quality of service of a distributed system, and provides application-level mobility support for user mobility. This approachIt has the default assumptiones of that user locations are sensing known and that the network interface being is preprovideddefined virtually. This one architecture also concentrates on the adaptingation of applications to the ubiquitous and heterogeneous system environments. B, but it is relatively little unconcerned with using the user contexts and profiles for userto support mobility.
The A different middleware, described by [6], address QoS and scalability issue by other approach for the quality of service and scalability issues is software architecture-based adaptation for providing the tools to managhandleing varying dynamic resources, user mobility, dynamicchanging user needs, and system faults. [69]. It provides an This approach illustrates adaptation framework with with three layers: the task layer, model layer, and runtime layer. While this model is sound from the architectural point of view, it still requires additional features for assurance of quality of service in each layer for user mobility.
The Service Reference Model of Working Group 2 inThe Wireless World Research Forum (WWRF) provides a Service Reference Model describinges service semantics, generic service elements, and a service platform for solving the binding, security and privacy issues. As tThe early specifications specification is promising, but are in the early stages, they still lacks functional details between each component [172].
Most of the related work is generally focuses on one individual problem as associatedrelevant withto one particular domain, but many of these problems cross domains, and new problems arise when you attempt to address the collective issues. Therefore each solution applies to one particular domain.lack of relationship between other issues and also lack of method for providing proper service of people’s needs. However, aANomadic pervasive computing is one of these cross-domain problems. We identify the domains and analyze specific issue that must be addressed to provide continuous service across different smart spaces.
As seamless communication service for user mobility is at the beginning stage, additional work is required to link seamless service to the ubiquitous system.
3. Challenges of Nomadic Pervasive Space
Assuming there are isolated islands of smart spaces lying around, and people or objects are moveing about from one to another. When people roam around between these smart spaces, they would like to receive as much assistance as feasible on the road as if they stay at the comfort of their own smart house. However, tThe realization of the seamless continuous service provisioning from smart spaces requires demands the following challenges to be addressed met first. SCENARIO
A. Binding issues:
1. Separating Goals and Means. :
It is crucial to separate goals and means in nomadic pervasive spaces. (because…). Consider a medicine reminder service. For instance, bBroadcasting messages in on the speaker stereo system at home can certainlyis an excellent means to remind prompt the residentoccupant to take medicine on time. But it may not be the most feasible modality if when the person is shopping in at the supermarket. ,Here, an alternative mean, such as text messaging to the c the userellular phone would be much more appropriate. Some aspects of this issue hasve been widely researched related in regards to context awareness. In nomadic pervasive spacecomputing, the problem has been intensified because of the difference in location and potentiallydifferent spaces may have drastically difference different in the availability capabilities for of the context acquisition and behavior delivery.
2. Deciding whereLevel of which Bbinding take placeoccurs
When a user arrives in a new smart space, at which level are the user’s preferences bound to the space’s configuration?
Upon arrival to the new smart space, how do binding take place between users’ preference and the local system? On one hand, tOnehe most intuitive ideas is to bind at the device level. With this approach, each device is associated with particular domains, and services bind to the devices that support their needed domains. to defined context, with which we connect the mean of actuations based on the availability of the devices present, fFor instance, a climate the control service, which affects the “temperature” domain, of room temperature can be connected bound to devices such asto an air conditioner, electric heater, or ceiling fanblinds. On the other hand, th
Another option is toere are other ideas, such as binding bind at the user profile level,. in whichHere, the user’s’ preferences are adapted and converted to conform to the protocols of the enew spacexisting system understood by the local smart space
There are of course other levels at which binding could occur. Research in nomadic pervasive computing will involve identifying the factors that affect where binding should occur for particular architectures or applications.. As another example, we can even bind at the context graph level,
in which the entire relations between contexts and users’ preference and expected actions have been migrated from one smart space to the other, it would be similar to running a user-specific virtual machine to provide the user with the customized services.
3. CompatibilityInteroperability and Compatibility:
Device descriptions, Service/Application descriptions and user preferences are some of the crucial information to make a smart space click. The syntactic and semantic grammars have to be well defined for the entire smart space to work properly. On top of this, the difference in programming models may also cause incompatibility. Many researchers have looked into extensive use of ontology and standardization of service description, but no definite conclusion has been reached yet. While this issue exists in single-location smart space, it is relatively easier to ensure compatibility of entities under the same roof than roaming around to different locations between systems, possibly internationally.
4. Ownership of Space and Authentication of Entities
When users and devices move in and out of smart spaces freely, issues related to authentication and ownership become central. In the private setting of a smart home, the owner of the house should have the full control of the dwelling. Any device and service brought in or approved by the owner can be trusted and integrated into the environment. In the semi-private setting, such as paying visits to a friend’s smart home, some sort of rules have to be established to accommodate the visitors’ needs while respect the control and preference of the owner. For instance, it would be more hospital to allow the visitor to setup the room temperature of the guest room, but not the master suite. In the public setting, such as a trip to the supermarket, a more restricted model may be adapted to better serve the large number of occupants to the public place, while still offer important services, such as patching through shopping list or medicine reminder to users’ handheld device or smart phone.
The system has to be able to recognize the boundary as to which services are legitimate and allowed, while others may cause adverse effects to the occupants or the interest of general public. How to characterize each person, device and service, and authenticate them and identify their ownership is one of the key concerns for nomadic pervasive space.
B. Quality of Service:
1. Conflict resolution: Priority and Scheduling
Conflict resolution can be regarded as another facet of the authentication and ownership issue, although it is crucial in their own right. How do we decide which of the conflicting commands issued by different applications should overwrite all the others? How do we decide the priority of conflicting preferences of users? When including time dimension in the consideration, how do we schedule the execution of the commands so as to handle emergency situations promptly and preserve the quality of service in general.