The Elastic Processing of Data Streams in Cloud Environments:
A Systematic Mapping Study
Floriment Klinaku, Michael Zigldrum, Markus Frank, and Steffen Becker
Reliable Software Systems, University of Stuttgart, Germany
Keywords:
Elastic, Processing, Data Streams, Cloud.
Abstract:
Ongoing efforts exist to exploit cloud elasticity for processing efficiently data streams generated by a variety
of data sources. To contribute to these efforts an overview of existing research is required. To the best of our
knowledge, a systematic overview of the field is missing. To fill this gap, we conduct a Systematic Literature
Map (SLM). This way we offer a high-level overview of the literature on elastic data stream processing.
We search four databases, evaluate 564 publications and identify 100 relevant publications. The identified
publications show that the majority of work is validated research through proofs-of-concept and very few
through case studies, surveys and field experiments. There are several frameworks, approaches and tools
proposed, but, fewer metrics, models and processes.
1 INTRODUCTION
Motivation. Billions of data sources (i.e., sensors,
devices, smartphones or things
1
) are connected to the
Internet. They continuously produce data of different
structure at variable rates. The variability in arrival
rates and structure places a challenge on software sys-
tems that are responsible for processing such streams
in an efficient and timely manner. The rise of the
cloud computing paradigm (Mell et al., 2011) made it
possible for businesses and customers to acquire com-
puting resources according to their needs. Moreover,
there exist ongoing efforts on designing processes
and solutions (e.g., auto-scalers) which autonomously
handle the provisioning of resources (Chen et al.,
2018) in the presence of changing conditions and de-
mands.
Problem. One class of software systems that tend to
exploit cloud elasticity are stream processing systems.
These systems perform computations on events that
continuously arrive. They are required to efficiently
compute low-latency queries on non-homogeneous
data while it arrives and in presence of variable arrival
rates. In literature these systems are also referred as
stream computing, complex event processing (CEP)
or data stream management system (DSMS) (Hummer
et al., 2013).
The importance of these systems will continue to
increase based on predicted increase of sensors and
IoT devices in the next years. Thus, it is important
1
Internet of Things vocabulary
for researchers and engineers to have an overview of
existing and missing work in state-of-the-art research
for elastic and efficient processing of data streams. To
the best of our knowledge an overview is missing.
Solution. To provide an overview of existing research
on elastic stream processing we conduct a System-
atic Literature Map (SLM). Through this work we
make the following contributions: 1. We systemati-
cally present work of researchers which tackles as-
pects in exploiting cloud elasticity for data stream
processing applications. 2. We classify publications
based on two facets: research and contribution and
identify areas which need more attention in the future.
3. We discuss publications in areas where more work
is required and provide researchers high-level aspects
of their work.
Results. A set of one hundred publications constitute
the effort of researchers tackling different aspects of
elasticity in the context of stream processing. There
are gaps in evaluation and opinion type of research.
With respect to contribution type there are several
frameworks and approaches found in the literature but
less work is done in metrics, models and discussions.
Structure. The rest of the work is structured as fol-
low: Section 1 presents details on the method. Sec-
tion 3 presents the results; Section 4 discusses threats
to validity; Section 5 highlights related work; finally
Section 6 concludes the work and highlights plans for
the future.
316
Klinaku, F., Zigldrum, M., Frank, M. and Becker, S.
The Elastic Processing of Data Streams in Cloud Environments: A Systematic Mapping Study.
DOI: 10.5220/0007708503160323
In Proceedings of the 9th International Conference on Cloud Computing and Services Science (CLOSER 2019), pages 316-323
ISBN: 978-989-758-365-0
Copyright
c
2019 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
Selectreference
papers
Conductsearch Screenpapers
Designclassification
scheme&classify
Extractdataand
performmapping
Constructsearchqueries
100
"Snowballing"
4
16 564
100
Search,filterandvote
Research
type
Contribution
type
Classification
Systematic
map
Map
Search
queries
Figure 1: The overall study process.
2 RESEARCH PROCESS
To ensure reproducible and valid results we adhere to
the process and guidelines described by Kuhrmann et
al. (Kuhrmann et al., 2017) and by Petersen et al. (Pe-
tersen et al., 2008). In this section, we describe unique
characteristics of the research process and avoid re-
dundant guidance on how to conduct mapping studies.
For details left out, readers should follow the process
by Petersen et. al. To discover the state-of-the-art re-
search and identify gaps in the literature we chose to
quantify the body of work through a Systematic Lit-
erature Map (SLM). We define the research scope as
follows:
Context. Processing data streams in cloud environ-
ments
Peculiarity. Elasticity
We design all the activities in the study process ac-
cording to the Context and Peculiarity To determine
how well the defined peculiarity is exploited in the
defined context we pose two research questions:
RQ
1
: How large is the body of work that tackles
cloud elasticity related peculiarities in the field
of data stream processing?
RQ
2
: What type of research is required for the fu-
ture?
Figure 1 depicts the overall step-by-step process
which led to the results presented in this paper. From
left to right: we start off with a set of 4 reference
publications (Cervino et al., 2012a; Gedik et al.,
2014; Hochreiner et al., 2016; Abadi et al., 2005),
which objectively belong to the publication space that
is mapped. Through snowballing (Kuhrmann et al.,
2017) we extended the initial set of relevant work to
16 publications. We use this set of publications to ob-
tain the search queries iteratively. The criteria for con-
structing successful search queries was to obtain the
initial set of 16 publications when searching across
the data sources. Since there were resources where
the abstract of the work was not given we designed
two queries: one for searching on titles and one for
Table 1: Final search queries.
(a) Title search string
(elastic or auto?scale or borealis) and (stream or
streaming) and (processing or system)
(b) Abstract search string
(elastic or auto?scale or borealis) and (data or
cloud or event or query) and (stream) and (pro-
cessing or computing or system)
abstracts as shown in Table 1.
We search in four prominent databases: ACM,
IEEE, Springer, Science Direct. We decided to se-
lect these sources based on their focus on computer
science and are also used in other mapping studies
related to software engineering (e.g., (Ingibergsson
et al., 2015)). Moreover, considering the search terms
for our mapping study like ”elasticity” or ”streams”—
which are very popular in other fields like physics
or economics—we decide not to search in a meta-
search database like Google Scholar. After searching
the selected databases and filtering out duplicates and
publications not relevant to our context (e.g., video-
streaming), we obtain a set of 564 publications. To
further evaluate the found papers we follow the voting
process proposed in (Kuhrmann et al., 2017), where
reviewers vote independently and then in a workshop
reviewers discuss publications which received differ-
ent votes. Finally, after the voting process, there were
100 publications left to be classified and mapped.
To get a better overview of the publication space,
we classify the papers based on research and con-
tribution type. For the first classification facet
we follow the proposed scheme by Wieringa et al.
(Wieringa et al., 2006). For the later—the con-
tribution type—we created our own scheme. For
the research type, there are seven classes, namely
Evaluation, Solution Proposal, Validation, Philosoh-
pical, Opinion, Personal Experience, Review, and
The Elastic Processing of Data Streams in Cloud Environments: A Systematic Mapping Study
317
Table 2: Research type classification.
Name Abbr Description - The work ...
(a) Research type classification facet
Evaluation EVAL “results in new knowledge of causal relationships among phenomena. Causal properties are studied empirically, such as
by case study, field study, field experiment, survey, etc.”
Solution
Proposal
SOL “... proposes a solution ... and argues for its relevance, without a fullblown validation. A proof-of-concept may be offered
by means of a small example, a sound argument, or by some other means.” (Wieringa et al., 2006)
Validation VAL “... proposes a solution ... and argues for its relevance, without a fullblown validation. A proof-of-concept may be offered
by means of a small example, a sound argument, or by some other means.” (Wieringa et al., 2006)
Philosophical PHIL “... sketches a new way of looking at things, a new conceptual framework, etc.” (Wieringa et al., 2006)
Opinion OP “... contain[s] the author’s opinion about what is wrong or good about something, how we should do something, etc.”
(Wieringa et al., 2006)
Personal Expe-
rience
PERS “... will often come from industry practitioners or from researchers who have used their tools in practice, and the experi-
ence will be reported without a discussion of research methods. The evidence presented in the paper can be anecdotal.”
(Wieringa et al., 2006)
Review and
Summary
REV “... where author/authors is/are reviewing or summarizing the evolution of an area of research in a historical fashion.”
(Wieringa et al., 2006)
(b) Contribution type classification facet
Metric MET ... contributes to forming comparable and meaningful metrics. The work presents a benchmark tool and provides metrics
with which to compare.
Model MOD ... provides models which aid to verify approaches, algorithms or other research.
Process PROC ... describes a process or processes to reach a certain goal.
Discussion DISC ... conducts a literature review or discusses and outlines challenges based on already existing research papers.
Approach/ Al-
gorithm
APPR ... outlines a new approach to a known problem, or provides an algorithm which solves an existing problem.
Framework FRMW ... provides a complete working framework which tackles several challenges. Frameworks do not have to be fully imple-
mented or validated, as the contribution to the research field does not change based on its status.
Tool TL Contrary to a framework, work in this category contributes to a smaller part of a complete system or adds to an existing
framework. Tools are distinct from Approaches/Algorithms, as they are either tailored to a specific framework or are
self-contained solutions, which combine multiple algorithms into one component.
Summary. For the contribution type there are six
claseses, namely Metric, Model, Process, Discussion,
Approach/Algorithm, Framework, and Tool. Both are
shown in Table 2 part (a) and (b).
3 RESULTS
In order to get an overview of research with respect
to the utilization of cloud elasticity for efficient data
stream processing we examine the number of papers
that tackle this issue. We conduct a systematic map-
ping study to obtain the set of papers that contribute
to the defined research context and peculiarity.
Generally speaking, the 100 publications (Table
3
2
) constitute a valuable information for researchers.
Moreover, the data points highlight the lack of work
in several types both for the contribution facet as well
as the research facet. These gaps could constitute fu-
ture research directions. The rest of this section re-
2
Due to space reasons Table 3 contains selected
publications, to see the full list of publications visit
http://klinakuf.github.io/ms-elastic-datastreams
ports on results and is divided in two subsections. The
two subsections report on the data according to the
designed research questions.
RQ1: How large is the body of work that tack-
les cloud elasticity related pecularities in the big
data stream processing context? A set of one hun-
dred papers constitutes the attempt of researchers to
tackle different challenges in designing elastic and ef-
ficient big data stream processing applications. As
seen in Figure 2 the first relevant publication came
out in 2000. Then in 2012, the number of publica-
tions rose to a peak with 22 papers in 2015. After the
peak, in the following two years, the number of pub-
lications fell. It is important to note that the data for
the year 2018 is not complete, as we conduct the data
search and export in August of 2018.
RQ2: What type of research is required for the
future? As Figure 3 depicts, the biggest gaps are
in evaluations and opinions. Evaluation work need
to evaluate the research in practice whereas opinions
are usually formed after investigation of multiple dif-
ferent approaches. Both research types have a low
number of works. Also other research categories—
CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science
318
Table 3: Selected Publications.
RT CT Publications Num
VAL FRMW (Gkolemis et al., 2017; Katsipoulakis et al., 2015; Wu and Tan, 2015; Gedik et al., 2018; Madsen and Zhou, 2013) 5/31
APPR (Cammert et al., 2008; Kombi et al., 2017; Isert and Schwan, 2000; Cervino et al., 2012b; Vasconcelos et al., 2016) 5/18
TL (Wu et al., 2017; Heinze et al., 2015; Zhang et al., 2013; Cardellini et al., 2018; Bellavista et al., 2013) 5/16
MOD (Mencagli, 2016; Heinze et al., 2014b; Nguyen et al., 2015; Lin et al., 2015; Qanbari et al., 2015) 5
SOL FRMW (Vakali et al., 2016; Zhang et al., 2015; Bhandari, 2012; Chen et al., 2014; Zacheilas et al., 2016) 5/8
APPR (Das et al., 2014; Vu et al., 2010; Heinze, 2011; Humayoo et al., 2014) 4
DISC (Ahmed et al., 2016; Reale et al., 2014; Martin et al., 2014) 3
TL (Martins et al., 2014; HoseinyFarahabady et al., 2017) 2
MOD (Imai et al., 2016; Vulpe and Frincu, 2017) 2
PROC (Heinze et al., 2014c) 1
MET (Vorona et al., 2014) 1
REV DISC (Heinze et al., 2014a; de Assuncao et al., 2018; Hummer et al., 2013) 3
PHIL DISC (Eyers et al., 2012; Sun et al., 2015) 2
FRMW (Hochreiner et al., 2015) 1
APPR (Bustamante et al., 2001) 1
OP PROC (Truong et al., 2016) 1
EVAL FRMW (Chun et al., 2013) 1
2
1
2
1
2
8
22
15
11
15
3
2
88
0
5
10
15
20
25
2000 2005 2010 2015
Year
Publications
Figure 2: Number of publications per year.
excluding validation research—have a low number
of publications. Especially summaries in which this
study contributes to. With respect to contribution type
it is evident that the majority of publications is about
frameworks, approaches and tools and less about pro-
cesses, models and metrics.
This study helps to capture state-of-the-art re-
search in utilising elasticity for processing data
streams in cloud environments. Even though the num-
ber of publications in the last years of the graph de-
clined, the area can be considered an active area of re-
search with 11 publications in 2017 and eight publica-
tions until August 2018. The curve showing the num-
ber of publications fits the Gartner curve that shows
the adoption and maturity of an emerging technology
(Linden and Fenn, 2003). The Gartner curve assumes
that most technologies and concepts will progress
through the pattern of overenthusiasm and disillusion-
ment. One could judge the field as immature, passing
the overenthusiastic phase in 2012 and entering the
growth phase with a steady number of publications.
The study shows that the most significant gaps are
in evaluations and opinions. The found gaps rein-
force the judgment about the maturity of the field.
Evaluation work needs to evaluate the research in
practice whereas opinions are usually formed after
investigating different approaches. Both research
types have a low number of works. Also other re-
search categories—besides validation research—have
a small number of publications. Especially sum-
maries in which this study contributes to. Regard-
ing contribution type, it is evident that the majority
of publications is about frameworks, approaches and
tools, and less about processes, models, and metrics.
Processes. Concerning processes, (Truong et al.,
2016) emphasize the difficulty in engineering IoT
cloud platforms due to the lack of tools to test and
evaluate complex designs. For elasticity, they split
requirements in two parts: in analytical and control.
For the first, it is essential for developers and other
roles to have an end-to-end view on behavioral limits
(emergent behavior) so to enable proper roles in the
team to refine the software and improve control strate-
gies. For the second—elasticity control—they define
the granularity of control for different parts of the IoT
system (e.g., the data reading frequency of sensors or
The Elastic Processing of Data Streams in Cloud Environments: A Systematic Mapping Study
319
5
16
18
311
1
2
2
1
3
4
8
2
1
1
1
3
Approach
Discussion
Framework
Metric
Model
Process
Tool
Evaluation Opinion Philosophical Solution Summary Validation
Research Type
Contribution Type
Figure 3: Publications mapped to research and contribution type.
adding/removing virtual machines).
Metrics and Benchmarks. (Vorona et al., 2014)
contribute to the identified gap by benchmarking an-
alytical platforms for answering complex business
queries while leveraging the elastic infrastructure of
the cloud. Moreover, they propose two metrics scal-
ing overhead and elasticity overhead. The first rep-
resents the time wasted while the system is stabiliz-
ing after a provisioning action whereas the second—
elasticity overhead—represents the time lost because
of sub-optimal scaling decisions. It is of importance
to notice that these elasticity metrics differ when com-
pared to application agnostic elasticity metrics like
mean time to quality repair (MTTQR) proposed in
(Lehrig et al., 2015).
Models. Our study reveals various kind of mod-
elling approaches to enable elasticity for data stream
processing systems. These kinds include application-
agnostic performance modelling solutions (Imai et al.,
2016), game-theoretic approaches (Mencagli, 2016),
models which estimates the latency spike created by
a set of operator movements (Heinze et al., 2014b)
used then to built latency aware elastic placement op-
erator algorithm, models for data assets in the context
of Data-as-a-Service (DaaS) (Nguyen et al., 2015)
where providers could handle different quality re-
quirements for results.
4 THREATS TO VALIDITY
In this section we describe threats to the following
classes of validity as presented by (Claes et al., 2000):
(1) conclusion validity, (2) internal validity and (3)
construct validity.
Conclusion Validity. The choice of data sources
constitutes one threat to the validity of the conclusion
for the identified gaps. The study might potentially
miss relevant publications that exist on other sources.
Internal Validity. The classification of publica-
tions is another threat to the internal validity of the
study. The classification might be impacted by human
error both from authors of this study when classifying
known as judgmental error (Petersen et al., 2008), and
authors of the relevant publications. Potential errors
on both sides could lead to incorrect classification.
Construct Validity. The conclusion of the map-
ping study about uncovered contribution types, e.g.,
metrics or processes, is subject to the threat that
classes of the contribution classification facet may
subsume each other, e.g., a publication classified as an
approach may also contribute with metrics for evalu-
ating the approach. However, we argue that this does
not impact the identified gaps since there is a low
number of publications which have these classes as
their primary contribution.
CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science
320
5 RELATED WORK
Our mapping study revealed other works which con-
tribute to summarizing the problem area. In a recent
survey, Assunc¸
˜
ao et al. (de Assuncao et al., 2018)
discuss challenges, solutions and techniques for elas-
tic and efficient data stream processing. Their study
shows problems at a finer granularity whereas our re-
search gives an overview at a higher level and identi-
fies gaps for future research directions.
Heinze et al. (Heinze et al., 2014a) present open
challenges for next generations of data stream pro-
cessing systems. One aspect they foresee as important
is advanced elasticity where they argue that current
elasticity strategy optimize utilization of the system
only and other metrics like latency or bandwidth are
less considered.
In a similar direction goes also the work of Hum-
mer et al. (Hummer et al., 2013) where they pinpoint
one challenge which also our study reflects: how to
define service-level objectives in the best way and
how are metrics that indicate the elasticity of a data
stream processing system related to parameters that
reflect the quality guarantees for a tenant in a multi-
tenant data stream as a service scenario.
6 CONCLUSION
In this paper, we present the results of a systematic
literature map on elastic processing of data streams.
We search 4 databases, evaluate 564 publications and
identify 100 relevant publications which tackle differ-
ent elasticity aspects of data stream processing. The
study aids researchers in the field to get a high-level
overview of contributions and forms a basis for a de-
tailed study. Moreover, the visual map based on con-
tribution and research type helps practitioners to iden-
tify work which is relevant for their specific problems
in the area of elastic data stream processing.
For future work, we plan a more in-depth study
of the publication set. We plan to extract information
concerning contribution areas that show a low number
of publications, e.g., metrics, models and processes.
One possible way is to analyse publications that are
contributing to frameworks or tools. From these pub-
lications, one could extract and compare information
about metrics, models and processes they have used.
This in-depth study would benefit the body of work
with a better overview of the identified gaps.
REFERENCES
Abadi, D. J., Ahmad, Y., Balazinska, M., C¸ etintemel, U.,
Cherniack, M., Hwang, J.-H., Lindner, W., Maskey,
A., Rasin, A., Ryvkina, E., Tatbul, N., Xing, Y., and
Zdonik, S. B. (2005). The Design of the Borealis
Stream Processing Engine. Cidr.
Ahmed, T. M., Zulkernine, F. H., and Cordy, J. R. (2016).
Proactive auto-scaling of resources for stream pro-
cessing engines in the cloud. In Proc. of the 26th An-
nual Intl. Conf. on Comp. Science and Software Engi-
neering.
Bellavista, P., Corradi, A., Kotoulas, S., and Reale, A.
(2013). Dynamic datacenter resource provisioning for
high-performance distributed stream processing with
adaptive fault-tolerance. In Proc. Demo & Poster
Track of ACM/IFIP/USENIX Intl. Middleware Conf.
Bhandari, S. (2012). An adaptive event stream processing
environment. In Proc. of the on SIGMOD/PODS PhD
Symp.
Bustamante, F. E., Eisenhauer, G., and Schwan, K. (2001).
The active streams approach to adaptive distributed
systems. In High Performance Distributed Comput-
ing. Proc. 10th IEEE Intl. Symp. on.
Cammert, M., Kramer, J., Seeger, B., and Vaupel, S. (2008).
A cost-based approach to adaptive resource manage-
ment in data stream systems. IEEE Transactions on
Knowledge and Data Eng., 20(2).
Cardellini, V., Presti, F. L., Nardelli, M., and Russo, G. R.
(2018). Decentralized self-adaptation for elastic data
stream processing. Future Generation Comp. Systems.
Cervino, J., Kalyvianaki, E., Salvachua, J., and Pietzuch, P.
(2012a). Adaptive provisioning of stream processing
systems in the cloud. In IEEE 28th Intl. Conf. on Data
Engineering Ws.
Cervino, J., Kalyvianaki, E., Salvachua, J., and Pietzuch,
P. (2012b). Adaptive provisioning of stream process-
ing systems in the cloud. In Data Engineering Ws.s
(ICDEW), IEEE 28th Intl. Conf. on.
Chen, T., Bahsoon, R., and Yao, X. (2018). A survey and
taxonomy of self-aware and self-adaptive cloud au-
toscaling systems. ACM Comput. Surv.
Chen, W.-s., Chen, Y.-J., and Wu, S.-y. (2014). Dynamic
aggregate: An elastic framework for qos-aware dis-
tributed processing of rfid data on enterprise hierar-
chy. IEEE Transactions on Parallel and Distributed
Systems.
Chun, B.-G., Condie, T., Curino, C., Douglas, C., Matu-
sevych, S., Myers, B., Narayanamurthy, S., Ramakr-
ishnan, R., Rao, S., Rosen, J., et al. (2013). Reef:
Retainable evaluator execution framework. Proc. of
the VLDB Endowment.
Claes, W., Per, R., Martin, H., Magnus, C., Bj
¨
orn, R.,
and Wessl
´
en, A. (2000). Experimentation in soft-
ware engineering: an introduction. Online Available:
http://books. google. com/books.
Das, T., Zhong, Y., Stoica, I., and Shenker, S. (2014). Adap-
tive stream processing using dynamic batch sizing. In
Proc. of the ACM Symp. on Cloud Computing.
de Assuncao, M. D., da Silva Veith, A., and Buyya, R.
(2018). Distributed data stream processing and edge
The Elastic Processing of Data Streams in Cloud Environments: A Systematic Mapping Study
321
computing: A survey on resource elasticity and future
directions. Journal of Network and Comp. Applica-
tions.
Eyers, D., Freudenreich, T., Margara, A., Frischbier, S.,
Pietzuch, P., and Eugster, P. (2012). Living in the
present: On-the-fly information processing in scalable
web architectures. In Proc. of the 2Nd Intl. Ws. on
Cloud Computing Platforms.
Gedik, B. et al. (2018). C-stream: A co-routine-based elas-
tic stream processing engine. ACM Transactions on
Parallel Computing.
Gedik, B., Schneider, S., Hirzel, M., and Wu, K. L. (2014).
Elastic scaling for data stream processing. IEEE
Transactions on Parallel and Distributed Systems.
Gkolemis, E., Doka, K., and Koziris, N. (2017). Automatic
scaling of resources in a storm topology. In Intl. Ws.
on Algorithmic Aspects of Cloud Computing.
Heinze, T. (2011). Elastic complex event processing. In
Proc. of the 8th Middleware Doctoral Symp.
Heinze, T., Aniello, L., Querzoni, L., and Jerzak, Z.
(2014a). Cloud-based data stream processing. In Proc.
of the 8th ACM Intl. Conf. on Distributed Event-Based
Systems.
Heinze, T., Jerzak, Z., Hackenbroich, G., and Fetzer, C.
(2014b). Latency-aware elastic scaling for distributed
data stream processing systems. In Proc. of the 8th
ACM Intl. Conf. on Distributed Event-Based Systems.
Heinze, T., Pappalardo, V., Jerzak, Z., and Fetzer, C.
(2014c). Auto-scaling techniques for elastic data
stream processing. In Data Engineering Ws.s, IEEE
30th Intl. Conf. on.
Heinze, T., Zia, M., Krahn, R., Jerzak, Z., and Fetzer, C.
(2015). An adaptive replication scheme for elastic
data stream processing systems. In Proc. of the 9th
ACM Intl. Conf. on Distributed Event-Based Systems.
Hochreiner, C., Schulte, S., Dustdar, S., and Lecue, F.
(2015). Elastic stream processing for distributed envi-
ronments. IEEE Internet Computing.
Hochreiner, C., Vogler, M., Schulte, S., and Dustdar, S.
(2016). Elastic Stream Processing for the Internet of
Things. In IEEE 9th Intl. Conf. on Cloud Computing
(CLOUD).
HoseinyFarahabady, M., Lee, Y. C., Zomaya, A. Y., and
Tari, Z. (2017). A qos-aware resource allocation con-
troller for function as a service (faas) platform. In Intl.
Conf. on Service-Oriented Computing.
Humayoo, M., Zhai, Y., He, Y., Xu, B., and Wang, C.
(2014). Operator scale out using time utility function
in big data stream processing. In Intl. Conf. on Wire-
less Algorithms, Systems, and Applications.
Hummer, W., Satzger, B., and Dustdar, S. (2013). Elastic
stream processing in the cloud. Wiley Interdisciplinary
Reviews: Data Mining and Knowledge Discovery.
Imai, S., Patterson, S., and Varela, C. A. (2016). Cost-
efficient elastic stream processing using application-
agnostic performance prediction. In 16th IEEE/ACM
Intl. Symp. on Cluster, Cloud and Grid Computing.
Ingibergsson, J. T. M., Schultz, U. P., and Kuhrmann, M.
(2015). On the use of safety certification practices
in autonomous field robot software development: A
systematic mapping study. In Intl. Conf. on Product-
Focused Software Process Improvement.
Isert, C. and Schwan, K. (2000). Acds: Adapting computa-
tional data streams for high performance. In Parallel
and Distributed Processing Symp. IPDPS. Proc. 14th
Intl.
Katsipoulakis, N. R., Thoma, C., Gratta, E. A., Labrinidis,
A., Lee, A. J., and Chrysanthis, P. K. (2015). Ce-
storm: Confidential elastic processing of data streams.
In Proc. of the ACM SIGMOD Intl. Conf. on Manage-
ment of Data.
Kombi, R. K., Lumineau, N., and Lamarre, P. (2017).
A preventive auto-parallelization approach for elastic
stream processing. In Distributed Computing Systems,
IEEE 37th Intl. Conf. on.
Kuhrmann, M., Fern
´
andez, D. M., and Daneva, M. (2017).
On the pragmatic design of literature studies in soft-
ware engineering: an experience-based guideline.
Empirical software engineering, 22(6).
Lehrig, S., Eikerling, H., and Becker, S. (2015). Scalability,
elasticity, and efficiency in cloud computing: A sys-
tematic literature review of definitions and metrics. In
Proc. of the 11th Intl. ACM SIGSOFT Conf. on Qual-
ity of Software Architectures.
Lin, Q., Ooi, B. C., Wang, Z., and Yu, C. (2015). Scal-
able distributed stream join processing. In Proc. of the
ACM SIGMOD Intl. Conf. on Management of Data,
SIGMOD ’15.
Linden, A. and Fenn, J. (2003). Understanding gartner’s
hype cycles. Strategic Analysis Report N
o
R-20-1971.
Gartner, Inc.
Madsen, K. G. S. and Zhou, Y. (2013). Elastic mapreduce-
style processing of fast data. In Proc. of the 7th ACM
Intl. Conf. on Distributed event-based systems.
Martin, A., Silva, R., Brito, A., and Fetzer, C. (2014). Low
cost energy forecasting for smart grids using stream
mine 3g and amazon ec2. In Proc. of the IEEE/ACM
7th Intl. Conf. on Utility and Cloud Computing.
Martins, P., Abbasi, M., and Furtado, P. (2014). Audy:
Automatic dynamic least-weight balancing for stream
workloads scalability. In IEEE Intl. Congress on Big
Data.
Mell, P., Grance, T., et al. (2011). The nist definition of
cloud computing.
Mencagli, G. (2016). A game-theoretic approach for elastic
distributed data stream processing. ACM Transactions
on Autonomous and Adaptive Systems (TAAS).
Nguyen, T.-D., Truong, H.-L., Copil, G., Le, D.-H.,
Moldovan, D., and Dustdar, S. (2015). On developing
and operating of data elasticity management process.
In Intl. Conf. on Service-Oriented Computing.
Petersen, K., Feldt, R., Mujtaba, S., and Mattsson, M.
(2008). Systematic mapping studies in software en-
gineering. In EASE, volume 8.
Qanbari, S., Farivarmoheb, A., Fazlali, P., Mahdizadeh, S.,
and Dustdar, S. (2015). Telemetry for elastic data
(ted): Middleware for mapreduce job metering and
rating. In IEEE Trustcom/BigDataSE/ISPA.
Reale, A., Bellavista, P., Corradi, A., and Milano, M.
(2014). Evaluating cp techniques to plan dynamic re-
source provisioning in distributed stream processing.
CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science
322
In Intl. Conf. on AI and OR Techniques in Constriant
Programming for Combinatorial Optimization Prob-
lems.
Sun, D., Liu, C., and Ren, D. (2015). Prospects, challenges
and latest developments in designing a scalable big
data stream computing system. Intl. Journal of Wire-
less and Mobile Computing.
Truong, H.-L., Copil, G., Dustdar, S., Le, D.-H., Moldovan,
D., and Nastic, S. (2016). On engineering analytics for
elastic iot cloud platforms. In Intl. Conf. on Service-
Oriented Computing.
Vakali, A., Korosoglou, P., and Daoglou, P. (2016). A
multi-layer software architecture framework for adap-
tive real-time analytics. In IEEE Intl. Conf. on Big
Data.
Vasconcelos, R. O., Vasconcelos, I., and Endler, M. (2016).
Dynamic and coordinated software reconfiguration in
distributed data stream systems. Journal of Internet
Services and Applications.
Vorona, D., Funke, F., Kemper, A., and Neumann, T.
(2014). Benchmarking elastic query processing on big
data. In Ws. on Big Data Bench.
Vu, D., Kalogeraki, V., and Drougas, Y. (2010). Efficient
stream processing in the cloud. In Intl. Conf. on Het-
erogeneous Networking for Quality, Reliability, Secu-
rity and Robustness.
Vulpe, A. and Frincu, M. (2017). Scheduling data stream
jobs on distributed systems with background load. In
17th IEEE/ACM Intl. Symp. on Cluster, Cloud and
Grid Computing (CCGRID).
Wieringa, R., Maiden, N., Mead, N., and Rolland, C.
(2006). Requirements engineering paper classification
and evaluation criteria: A proposal and a discussion.
Requirements Engineering, 11(1).
Wu, R., Liu, B., Chen, Y., Blasch, E., Ling, H., and Chen, G.
(2017). A container-based elastic cloud architecture
for pseudo real-time exploitation of wide area motion
imagery (wami) stream. Journal of Signal Processing
Systems.
Wu, Y. and Tan, K.-L. (2015). Chronostream: Elastic state-
ful stream computation in the cloud. In IEEE 31st Intl.
Conf. on Data Eng.
Zacheilas, N., Zygouras, N., Panagiotou, N., Kalogeraki, V.,
and Gunopulos, D. (2016). Dynamic load balancing
techniques for distributed complex event processing
systems. In Distributed Applications and Interopera-
ble Systems.
Zhang, F., Cao, J., Khan, S. U., Li, K., and Hwang, K.
(2015). A task-level adaptive mapreduce framework
for real-time streaming data in healthcare applica-
tions. Future Generation Comp. Systems.
Zhang, Z., Shu, H., Chong, Z., Lu, H., and Yang, Y. (2013).
C-cube: Elastic continuous clustering in the cloud. In
IEEE 29th Intl. Conf. on Data Engineering.
The Elastic Processing of Data Streams in Cloud Environments: A Systematic Mapping Study
323
