Management of context aware software resources deployed in a cloud environment for improving quality of mobile cloud services

International Journal of Distributed and Parallel Systems (IJDPS) Vol.5, No.5, September 2014
MANAGEMENT OF CONTEXT-AWARE SOFTWARE
RESOURCES DEPLOYED IN A CLOUD ENVIRONMENT
FOR IMPROVING QUALITY OF MOBILE CLOUD
SERVICES
Sohame Mohammadi1, Kamran Zamanifar2 and Sayed Mehran Sharafi3
Department of Computer Engineering, Najaf Abad Branch, Islamic Azad University,
Isfahan, Iran
ABSTRACT
In cloud computing environments, context information is continuously created by context providers and
consumed by the applications on mobile devices. An important characteristic of cloud-based context aware
services is meeting the service level agreements (SLAs) to deliver a certain quality of service (Qos), such as
guarantees on response time or price. The response time to a request of context-aware software is affected
by loading extensive context data from multiple resources on the chosen server. Therefore, the speed of
such software would be decreased during execution time. Hence, proper scheduling of such services is
indispensable because the customers are faced with time constraints. In this research, a new scheduling
algorithm for context aware services is proposed which is based on classifying similar context consumers
and dynamically scoring the requests to improve the performance of the server hosting highly-requested
context-aware software while reducing costs of cloud provider. The approach is evaluated via simulation
and comparison with gi-FIFO scheduling algorithm. Experimental results demonstrate the efficiency of the
proposed approach.
KEYWORDS
Cloud Computing, Context-Aware Computing, Context Management, SLA.
1. INTRODUCTION
Cloud computing is a distributed and parallel computing system that builds on the convergence
and advancement of several technologies, especially in utility and grid computing, autonomic
computing, hardware virtualization, service-oriented architecture, web services and the existing
relationship among them [1]. According to the NIST’s definition [2], Cloud Computing
consisting of five essential features namely on-demand self-service, broad network access,
resource pooling, rapid elasticity, and measured service. All services and resources in this
technology are provided as different services while cloud computing architecture is classified
based on a three-layer structure comprised of Infrastructure as a service (IaaS), platform as a
service (PaaS) and software as a service (SaaS) [1, 2].
Due to the improvement of mobile applications and emergence of the concept of cloud computing
in the last decade, cloud computing was introduced as a potential technology for mobile services
including context-aware services. Although most computing devices (e.g. mobile phones, PDAs,
laptops, tablets, etc.) in the cloud computing environment are in small sizes and can accompany
the users in all places, they face with resource constraints (e.g. battery life, cpu speed, storage,
DOI:10.5121/ijdps.2014.5501 1

and bandwidth). Therefore, resource constraints prevent quality of mobile services. Hence, the
data processing and storage would be carried out outside the mobile device in the cloud while
mobile device operates as a displaying platform [3].
A context-aware service continually collects the context of the environment and adapts its
operation to the context data. When a request is sent by a customer to the context-aware software
in the cloud computing distributed environment, the response to the request of context-aware
software is affected by loading extensive context data from multiple resources on the chosen
server [4, 5]. Therefore the performance of such software would be decreased during execution.
Moreover, in the cloud computing environment, the users of cloud services are grouped into
various classes according to the costs they have paid. The services are managed properly provided
that those of higher qualities are dedicated to the users who have paid more. Meeting service level
agreement and decreasing managing costs of resource are important Challenges for Cloud
providers. If cloud providers either fail to meet the objectives set in the agreements or provide
services of lower qualities to customers than what agreed upon, they would be penalized by
claiming them the costs or through recharging their accounts [1, 5], which results in reduction of
the total profit earned by the cloud providers.
In this study, a new scheduling algorithm is suggested relying on classifying similar context
queries and dynamic scoring of requests to enhance the efficiency of servers hosting highly
requested context-aware software. The proposed approach is based on the idea that
communications costs during execution of an application are regarded as overhead, if especially
communications are established for loading context information from far resources. The costs of
communications could be compensated by making use of multithread applications through
running proper requests with regard to the context data type which results in the improvement of
the efficiency of provided services. Also what might happen is that many users are looking for
similar context information. In the other words, lots of requests in queue are related to getting
access and processing other similar context information. Therefore, conscious scheduling
considerably reduces the indicator of mean response time and the costs of cloud providers.
The rest of the paper is organized as follows. The related works are discussed in the next section.
Sections 3 and 4 are dedicated to the background concepts and Problem formulation. The
proposed algorithm is represented in section 5. Section 6 describes the experimental results and
the overall evaluations. Finally, the conclusion remarks and states future direction of this research
are identified in Section 7.
2. RELATED WORK
In recent years, many job scheduling methods have been proposed in the grid and other
distributed environments. In [7], a comprehensive review of the scheduling algorithms in the grid
environment and their comparisons are presented. In contrast, due to the unique features of the
cloud computing environment, its scheduling problems are different. On the other hand, a new
generation of applications known as context-aware applications has emerged and cloud
infrastructure has been introduced to be a suitable technology for context-aware applications.
Context-aware job scheduling, management and obtaining required context data are the main
challenges for the context-aware services. A notable number of context provisioning and
management approaches have been proposed. Surveys of which have been published in order to
understand the features of existing systems for instance in [8-11]. In the following, a brief review
on some of the previous studies related to the present study is discussed.
Boloor et al have investigated a research regarding the scheduling and dynamic allocating of
requests to manage the context-aware software in the cloud distributed environment [4-6]. In this
2

study, the cloud distributed environment is modeled as several data centers. An attempt is made to
reduce the overall penalties charged to the cloud through proposing Data-aware Session-grained
Allocation with gi-FIFO Scheduling (DSAgS), a novel decentralized request management scheme
deployed in each of the geographically distributed datacenters, context cache replacement policies
to be applied at the time of loading new context data to replace the prior contexts when the cache
is full and considering the service level agreements aiming at saving time[4-6].
S. L. Kiani et al in the article [12] have proposed a new mechanism of cache management in the
cloud environment to manage the context-aware software. In this mechanism, according to
different types of context information, validity period and access patterns, the cache is divided
into two sections in a scalable way and the efficiency of the whole system is improved by
employing appropriate scheduling policies for each section. Although the proposed idea could be
considered an effective solution to improve the system, but the centered design and usage of a
cache in the context broker, which is main coordinating component in this system, leading to the
creation of bottleneck. Therefore, the architecture lacks scalability and location transparency.
Assuncao et al in the article [13] have resorted to the cloud computing infrastructure to overcome
the resource constraints in the mobile environment. Combining the context awareness and
adaptive job scheduling technologies, the study has proposed a context-aware scheduler. The
main objective of this model is resources utilization in the cloud computing environments to
improve the quality of services. Y. Zhu et al in [14] have proposed a framework for the logical
stream that organizes services in the cloud computing environment in a way that makes them
capable of adapting to the environment’s contextual changes and using information and resources
of the cloud computing during execution. The model has been implemented on Java platform.
Making use of context information to adapt services to the environment’s changes, sharing
resources and improvement of the utilization of the cloud resources, portability, flexibility and
interoperability of the services could be considered its principal features.
Badidi & Esmahi in [15] did their attempts to optimize scalability, interoperability and quality of
services. Therefore, a multi-attributes decision algorithm and a framework for preparation of the
context information were proposed in a way that they were based on the implementation of the
context-aware services on the cloud side, utilization of the context broker as a mediator between
context consumers and context services and employment of publish/ subscribe model. The
context providers in the publish/subscribe model expose their data items to the context-aware
applications to be read. If an application wants to subscribe to a context, it sends its subscription
request to the broker. The subscriptions have to be adapted to the data items when the context
information is changed. Therefore, the broker decides to send an alarm to the applications
subscribed to this context. Then the applications could take delivery of the published context
values [16]. Although the proposed idea in this study is to attract attentions, but no discussion is
made regarding the ways brokers could be used as mediators and it has merely remained as an
idea.
3. BACKGROUND
Some of the concepts related to the present study are defined in this section.
3.1. Cloud-based context services
Dey and Abowd [17] define context as “any information that can be used to characterize the
situation of an entity. An entity is a person, place, or object that is considered relevant to the
interaction between a user and an application, including the user and application themselves”.
Context includes two categories [17]: primary contexts are location, time, activity and identity,
which are considered as the most important context information. Secondary contexts involving
3

any other information that can be retrieved from the primary contexts such as weather, email, etc.
different types of contextual information have varying temporal validity durations and remaining
valid for a particular duration.
Context-aware services continuously adapt their state behavior with the context changes. Context
services are located on the clouds side, and context consumers (CxCs) look up and invoke the
services [18]. In cloud environments, CxCs are context data regarded as inputs to adapt their
behavior to the users’ current situation (e.g. context aware services). Context provider (CxP) is an
entity which provides context information. An CxP gathers raw data from context sources(e.g.,
sensors, networks, mobile phones, web services, etc), and convert them to meaningful
information, aggregation, modeling, reasoning and uploaded to the context broker(CxB).The CxB
acts as an intermediary to control context flow between CxPs and CxCs. Its main function is to
query resolution, event management, registration of available CxPs, provide routing and lookup
services [19, 20].
There are two cases of utilizing context information in cloud environment [18]: The first context
service is used to guide and provide improved service to the user. The second, context is used to
find fault and maintain the system stable and adapt their operations to change environmental
conditions.
3.2. Service level agreement
Cloud computing is a parallel and distributed computing systems that enable the deployment and
execution of applications and services on remote data centers with a pay-per-use business model
based on SLA over the internet. The SLA established through negotiation between the cloud
provider and the cloud consumer to ensure service quality. Negotiation strategy between the
cloud provider and consumer can investigate through the third party called broker. The SLA may
involve a variety of metrics to measure the quality of service requirements such as availability,
execution time, price, reliability, cpu usage and the storage used by the consumer[1,21].
One of the most applicable types of SLA is based on percentage which service Provider tries to
allocate resources to the customers in accordance with the agreed percentile during certain
periods of time. While in case of any failure regarding meeting the agreed level, they have to pay
fine [1].
The probabilistic SLA has the following parameters [6, 22]: Response time(R), which is duration
of the time it takes from when a request arrives at the cloud and leaves it. Response time
threshold (RT), the value of RT is the deadline of a request which must leave the cloud. Actual
conforming percentile: the percentile of requests of a particular user class which have met the
required response time. This value is controlled by the request management policy chosen by the
cloud provider. Desired conforming percentile: For different classes of users, this value is
determined through the negotiation between the cloud provider and the consumer. A penalty (P),
clause can be applied to the cloud provider, if the quality of service requirements is less than a
given value.
There are various penalty functions to calculate the cloud service providers’ penalties including
linear, exponential and stepwise functions. The main focus in this study takes stepwise penalty
function into accounts while bases requests on deadline to meet the satisfaction of the consumers.
4. STATEMENT OF THE PROBLEM
In recent years, increased use of mobile applications (e.g. context aware applications) and
emerging of cloud computing concept, cloud computing infrastructure has been introduced as a
potential technology for mobile-based environments in order to cope with resource constraints for
4

mobile users[3,13]. There are numerous servers in a distributed cloud environments, each one
hosting one or more context aware applications to provide services for multiple classes of CxCs.
When a request for a context aware service arrives at the datacenter through the internet, the
scheduling agent submits it to a suitable server. The job has to be scheduled by internal scheduler
at the end-server to execute it on resource and to send the answer back to the scheduling agent.
Context aware services utilize context to adapt themselves to their changing environment. The
response time of a request for a context aware application involves context provisioning, context
provider lookup times and time required to load context information on the chosen end server
[4,5,12]. Therefore, the speed of context-aware software would be reduced during execution
times in the cloud computing environment. Moreover, users of cloud services are grouped into
various classes according to the costs they have paid. Proper management of these services is met
provided that services of higher qualities are dedicated to those who have paid more [1, 5].
To solve the above mentioned problem, a study was carried out by Boloor et al regarding
scheduling and dynamic allocating of requests to manage context-aware software in the
distributed cloud environment [4- 6]. In this study, they propose Data-aware Session-grained
Allocation with gi-FIFO Scheduling (DSAgS) in a distributed cloud that is modeled as several
datacenters. The proposed scheduling scheme is based on the gi-FIFO scheduling algorithm
which has been mathematically proven in [23] to be the most appropriate for probabilistic SLAs
for a single server serving multiple class jobs. The scheduling policy they use to schedule
requests queued at each cloud server for the cloud provider hosting the context aware applications
and it could be described as follows: Firstly, choose the request class with the highest penalty;
then, amongst all the queued requests of the chosen class, choose one with maximum waiting
time resulting in a response time less than or equal to RT. If there is not such request, it will be
chosen the request with higher waiting period, consequently resulting in a response time greater
RT [4- 6].
Despite the fact that the proposed method in the present study could be considered a proper
solution regarding managing requests of the server hosting context-aware software in the cloud
environment but given the fact that the cloud servers are capable of responding to limited number
of requests at a specific moment and they should process and respond to high volumes of requests
based deadlines at that specific moment, so the number of customers losing deadline at that
operation would be increased while the efficiency of such highly-requested servers would be
drastically decreased and consequently cloud providers penalties could be increased. Therefore,
the present research aims at proposing a proper scheduling algorithm for the requests of context-aware
5
software in cloud computing environment with regard to the represented challenges.
5. THE PROPOSED CONTEXT-AWARE ALGORITHM
The decision making process of the scheduling algorithm to choose the next request to be
executed among requests queued at the end server is affected by classes’ penalties, deadline of
requests, context data storage type and differing patterns of updating context information in
dynamic and evolutionary environments. According to this problem, the context-aware
scheduling algorithm is engaged in multiple dimensions and complications related to finding a
way to estimate indefinite scores in a multidimensional space. Hence, a new approach of
scheduling is proposed for management of context-aware software requests.

6
Figure 1. Overall Architecture of the cloud based context provisioning system
The internal architecture of the context provisioning system containing some context providers on
a cloud infrastructure is depicted in figure 1. In a cloud computing environment, the context
customer including all computing devices of customers (e.g. mobile phone, PDAs, laptops and
other computational devices of cloud environment) send their requests for a particular context
type through the Internet[5, 6]. Users' requests are allocated to the cloud end servers based on
context-aware and dynamic scoring [5, 6]. Cloud servers might host one or more various context
providers, which each of them requiring for numbers of users’ requests. There are numerous
requests which call for similar context data types from such servers. The server hosting context-aware
software has to respond to lots of requests at any moment. Under these circumstances,
given the fact that the server is capable of responding to limited number of requests at a moment
and providing required context data for requests takes time, therefore the efficiency of the server
would be considerably reduced.
The problem is solved through proposing a solution which is based on classifying and dynamic
scoring of requests with regard to their user classes and the required context scope. Scope is a set
of closely related context parameters and context exchange unit that always requested, updated,
provided and stored at the same time [20, 24]. In this approach, each group of requests requiring
similar context types is considered a cluster in related scope while penalties for each cluster of
scope will be determined with regard to the number of requests related to the needed context data
and requests’ user classes.

7
Figure 2. The pseudo code of the proposed algorithm
Aiming at restricting the multidimensional space of request scoring, the gi-FIFO scheduling
algorithm is run from the beginning of running the proposed algorithm to estimate classes’
penalties and requests’ response times. Nevertheless, given that loading context data takes time, if
the required context information for running a request is not valid on a server at the time of
choosing the next request to be operated, it is indispensable that the required context data be
loaded on the server. During loading context data in order to compensate the communications
costs by computations, the scope should be chosen with the shortest expiry time on the server.
Regarding the updating pattern of the chosen scope, there are two main strategies behind the
proposed scheduling policy namely select the soonest expiring first (SE) for scope with short
update pattern and select the highest penalty (HP) score among the scopes that are valid on the
server for the scope with the long update pattern. Therefore, if the conditions are met with regard
to the loading time of context data, the requests related to the chosen scope would be
implemented based on gi-FIFO scheduling. figure 2 shows the pseudo-codes of the proposed
algorithm.
6. IMPLEMENTING AND EVALUATING
The proposed approach is simulated and evaluated by JAVA language programming. The
conformance levels, average response time and cloud providers’ penalties for executing requests
are the parameters analyzed. As for simulation, the simulation parameters of articles [5, 12] are
used. The input parameters for simulation are considered as follows: 1) 5 user classes, 2) 100
distinct sessions-id to assign the customer’s class, 3) customers are entered the system through
Poisson distribution with constant rate λ which is the number of customers at a time unit, 4) the
service processes are fixed 5) twelve different scopes are used in this simulation and the
parameters for each scope include: scope ID, the mean request processing time at context
provider, validity period of produced context data and types of scope updating patterns. The

values of scopes using from table 1, 6) the time of loading context data is considered three times
more than the average service rate for each request.
8
Table 1. Simulation Parameters [12]
CxP:ScopeID Processing time[ms] Validity[s] Category
CxP:1 70 60 Short
CxP:2 70 60 Short
CxP:3 80 80 Short
CxP:4 80 80 Short
CxP:5 90 180 Short
CxP:6 90 240 Short
CxP:7 70 360 Long
CxP:8 70 400 Long
CxP:9 80 600 Long
CxP:10 80 900 Long
CxP:11 90 1200 Long
CxP:12 90 1200 Long
After determining the simulation values, the users’ requests as the consumers of the context data
would be produced and the gi-FIFO scheduling algorithm and the proposed approach would be
implemented. Figure 3,4 show the plots of queues’ conformance levels vs. the number of user
requests. As the results show, regarding the gi-fifo scheduling policy, the growing rate of the
conformance levels decreases as the number of user requests increases and after it roughly
reaches zero. Thus experimental results reveal that the conformance levels obtained from the
proposed algorithm are higher than gi-FIFO scheduling algorithm for all user levels. As the
experimental results show, the proposed algorithm enjoys the best state at scheduling queue
requests relevant to the user class with the highest penalty while the worst state is at scheduling
those of the user class with the lowest penalty.
Figure 3.Comparison of conformance levels for the user class with the highest penalty

9
Figure 4. Comparison of conformance levels for the user class with the lowest penalty
Figure 5 shows the mean response time against the users’ requests. Given that a server is capable
of responding to a finite number of requests at a moment and it takes time to provide services
with the required context data, so according to the gi-FIFO scheduling policy, the percentile of
customers losing their deadlines at that operation would increase. Therefore, there would be a
significant increase in the mean response time. Nevertheless, In the proposed approach given that
requests are classified based on the required context information, they are scored dynamically and
the loading time of the context data is overlapped with the implementation of the requests, the
mean respond time is significantly reduced while increasing the total profit charged by the cloud
providers.
Figure 5. Comparison of average response time for proposed and gi-fifo schedules
For different classes of users are considered distinct penalty functions. The penalty functions for
multi-class jobs define the way penalty could be calculated. Our method for calculating the
penalties is based on the way presented by boloor [6] and is defined as follows[6]: If the effect of
delaying the recently arrived request causes the non-conformance to increase beyond that given in
the SLA, then the cloud provider is assigned a penalty of P$, else if the the non-conformance of
request is low, no penalty would be added to the total penalty. The value of current non-conformance
is calculated in below equation:
non-conformance = ((1 - cck) * Xk + 1)/( Xk + 1) (1)

Where Xk is total number of requests serviced of class k from the start of the observation interval
as measured by server d and cck is current conformance of class k in cloud as calculated by d [6].
Figure 6 shows the penalty of 4500 consumer queries incurred by the cloud provider with
different context data update pattern. The proposed policy is evaluated with different management
strategies regarding short validity (SV) scopes and long validity (LV) scopes in increment of 25%
[12]. Typical result in the figure shows that gi-FIFO policy results in the maximum penalties
charged. As for the approach with the shortest expiry time for scheduling as the duration of the
context load time, the lowest penalty is assigned when the scope distribution in the context-data
queries is biased towards SV scopes. The highest penalty value policy decreases the total penalty
charged to the cloud across all context queries. The above-mentioned strategy, however, is better
applied to LV context-data queries than SV context-data queries. The proposed scheduling makes
use of two different policies that are suitable for both short and long validity scopes distributed
context queries.
10
Figure 6: Comparison of total penalty incurred in gi-fifo and proposed schedules
7. CONCLUSIONS AND FUTURE WORK
In the cloud computing environment, since loading context data takes time, servers hosting
context-aware software are faced with the problem of low speed of providing services. Therefore,
the profit obtained decreases rapidly for cloud providers. In this research, a new scheduling
algorithm for context aware services is proposed which is based on dynamic scoring of the
requests to improve Qos. The proposed algorithm runs based on classifying requests in
accordance with the type of the required context data, dynamic scoring and overlapping loading
time of the context data with operating proper requests with regard to the storage type of the
context data and their differing updating patterns. Experimental results demonstrate the efficiency
of the proposed approach while highlight the necessity of managing the requests of the context-aware
software in the cloud environment.

In the future, we will investigate on modeling the probabilistic SLA globally in a distributed
cloud computing environment with numerous servers to stimulate developments in this
infrastructure.
REFERENCES
[1] R.Buyya, J.Broberg, and A.M.Goscinski, "Cloud Computing Principles and Paradigms," Wiley
11
Publishing, USA, 2011.
[2] P.Mell and T.Grance, "The NIST Definition of Cloud Computing", National Institute of Standards
and Technology, Information Technology Laboratory, Technical Report Version 15, 2011.
[3] L.Guan, X.Ke, M.Song, and J.Song, "A Survey of Research on Mobile Cloud Computing," in
Computer and Information Science (ICIS), 2011 IEEE/ACIS 10th International Conference on, pp.
387-392, 2011.
[4] K.Boloor, R.Chirkova, Y.Viniotis, and T.Salo, "Dynamic Request Allocation and Scheduling for
Context Aware Applications Subject to a Percentile Response Time SLA in a Distributed Cloud," in
Cloud Computing Technology and Science (CloudCom), 2010 IEEE Second International Conference
on, pp. 464-472, 2010.
[5] K.Boloor, R.Chirkova, T.Salo, and Y.Viniotis, "Management of SOA-Based Context-Aware
Applications Hosted in a Distributed Cloud Subject to Percentile Constraints," in Services Computing
(SCC), 2011 IEEE International Conference on, pp. 88-95, 2011.
[6] K.Boloor, "Management of soa-based, data-intensive applications deployed in a distributed cloud
subject to response time percentile service level agreements", Phd Thesis, North Carolina State
University, Raleigh, NC, USA, 2012.
[7] D.Amalarethinam, P.Muthulakshmi, "An Overview of the Scheduling Policies and Algorithms in Grid
Computing", International Journal of Research & Reviews in Computer Science, vol. 2, no. 2, pp.
280, 2011.
[8] M.Baldauf, S.Dustdar, and F. Rosenberg, "A survey on context-aware systems", International Journal
of Ad Hoc and Ubiquitous Computing, vol. 2, no. 4, pp. 263–277, 2007.
[9] H.Truong and S.Dustdar, "A survey on context-aware web service systems", International Journal of
Web Information Systems, vol. 5, no. 1, pp. 5–31, 2009.
[10] J.Hong, E.Suh, and S.Kim, "Context-aware systems: A literature review and classification", Expert
Systems with Applications, vol. 36, no. 4, pp. 8509–8522, 2009.
[11] M.Knappmeyer, S.L.Kiani, E.S.Reetz, N.Baker, and R.Tonjes, "Survey of Context Provisioning
Middleware," Communications Surveys & Tutorials, IEEE, vol. 15, pp. 1492-1519, 2013.
[12] S.L.Kiani, A.Anjum, K.Munir, R.McClatchey, and N. Antonopoulos, "Context Caches in the Clouds,"
Journal of Cloud Computing: Advances Systems and Applications, 2012.
[13] M.D.Assuncao, M.A.S.Netto, F.Koch, and S. Bianchi, "Context-Aware Job Scheduling for Cloud
Computing Environments," in Utility and Cloud Computing (UCC), 2012 IEEE Fifth International
Conference on, pp. 255-262, 2012.
[14] Y.Zhu, R.Y.Shtykh, and Q.Jin, "A Human-Centric Framework for Context-Aware Flowable Services
in Cloud Computing Environments," Journal of Information Sciences, 2012.
[15] E.Badidi, L.Esmahi, "A Cloud-Based Approach for Context Information Provisioning", Journal of
World of Computer Science and Information Technology Journal (WCSIT, pp. 63-70), 2011.
[16] E.Badidi, "A Publish/Subscribe Model for QoS-Aware Service Provisioning and Selection", Journal
of Computer Applications, vol. 26, 2011.
[17] G.Abowd, A.Dey, P. Brown, N. Davies, M.Smith, and P. Steggles, "Towards a Better Understanding
of Context and Context-Awareness", in Handheld and Ubiquitous Computing. vol. 1707, H.-W.
Gellersen, Ed., ed: Springer Berlin Heidelberg, 1999, pp. 304-307.
[18] H.Jung, S.Dong, "A Conceptual Framework for Provisioning Context-aware Mobile Cloud Services",
Cloud Computing (CLOUD), 2010 IEEE 3rd International Conference on, 2010.
[19] H.Vahdat-Nejad, K.Zamanifar and N.Nematbakhsh, "Towards a Better Understanding of Context
Aware Middleware: Survey and State of the Art", To be published, 2013.
[20] S.L. Kiani, A.Anjum, M. Knappmeyer, N.Bessis, and N. Antonopoulos, "Federated broker system for
pervasive context provisioning", Journal of Systems and Software, vol. 86, pp. 1107-1123, 2013.
[21] R.Rajavel and T.Mala, "Achieving Service Level Agreement in Cloud Environment Using Job
Prioritization in Hierarchical Scheduling", in Proceedings of the International Conference on
Information Systems Design and Intelligent Applications 2012 (INDIA 2012) held in Visakhapatnam,

India, January 2012. vol. 132, S. Satapathy, P. S. Avadhani, and A. Abraham, Eds., ed: Springer
Berlin Heidelberg, pp. 547-554, 2012.
[22] K.Boloor, R.Chirkova, T. Salo, and Y.Viniotis, "Analysis of Response Time Percentile Service Level
Agreements in SOA-Based Applications", in Global Telecommunications Conference (GLOBECOM
2011), 2011 IEEE, pp. 1-6, 2011.
[23] N.Agarwal and I.Viniotis, "Performance Space of a GI/G/1 Queueing System Under a Percentile Goal
Criterion," in Performance Modelling and Evaluation of ATM Networks, D. Kouvatsos, Ed., ed:
Springer US, pp. 474-484, 1995.
[24] M.Knappmeyer, S.L.Kiani, C.Fra, B.Moltchanov, and N.Baker, "ContextML: A light-weight context
representation and context management schema", in Wireless Pervasive Computing (ISWPC), 2010
5th IEEE International Symposium on, 2010, pp. 367-372.
12

Management of context aware software resources deployed in a cloud environment for improving quality of mobile cloud services

Recommended

More Related Content

What's hot (14)

Viewers also liked (19)

Similar to Management of context aware software resources deployed in a cloud environment for improving quality of mobile cloud services (20)

Recently uploaded (20)

Management of context aware software resources deployed in a cloud environment for improving quality of mobile cloud services