This document discusses quality management for service-based systems and cloud applications. It begins with quotes from Aristotle about how constantly performing certain actions can lead one to acquire particular qualities. It then discusses the need to manage how cloud services act in order to ensure quality. The document provides an overview of a proposed quality management architecture, including concepts like quality models, monitoring tools, and execution environments for analytics. It also reviews some existing quality models, monitoring techniques and tools, and cloud management platforms. Finally, it outlines next steps around designing and testing a complete quality management example.
ITALY - Visa Options for expats and digital nomads
Â
WP4-QoS Management in the Cloud
1. Quality Management in
Service-based Systems and
Cloud Applications
WP4 Quality Management and
Business Model Innovation
RELATE-ITN
Dr. Jose MarĂa Alvarez-RodrĂguez
Research Fellow, SEERC
Prague, 19-04-2013
2. âCloud-based services acquire a particular quality by
constantly acting a particular way... they become just by
performing just actions, temperate by performing
temperate actions, brave by performing brave actions.â
16/04/2013 Prague, Czech Republic #2
Aristotle
âMen acquire a particular quality by constantly acting a
particular way... you become just by performing just
actions, temperate by performing temperate
actions, brave by performing brave actions.â
âŠwe need to manage this particular way of acting!
Some time agoâŠ
4. 16/04/2013 Prague, Czech Republic #4
I need helpâŠ
I have a mobile application that needs a Geocoding service and
the response time must be in milliseconds.
âą More than 54 geocoding APIs
â How can I select the most suitable service?
â How can I compare different providers?
â How can I track the quality (response time) of the selected
service?
â âŠ
http://blog.programmableweb.com/2012/06/21/7-free-geocoding-apis-google-bing-yahoo-and-mapquest/
5. Context
ï¶ A growing offering of cloud services
ï¶ âŠmore complexity
ï¶âŠnew needs and requirements
ï¶ Cloud Management forâŠ
ï¶ Cloud models and types
ï¶ Track and control my third-party dependencies
ï¶ Context-aware quality dimensions/indicators/metrics
ï¶Security, storage, etc.
ï¶Subjective experience
ï¶ âŠto improve, optimize and accomplishâŠ
ï¶ efficiency, costs, SLAs, etc.
ï¶ by means of providing advanced services
ï¶ Analytics/Prediction/âŠ
#516/04/2013 Prague, Czech Republic
7. What is âQualityâ?
Classical view
Dimensions
ï¶ Tangibles
ï¶ Reliability
ï¶ Responsiveness
ï¶ Service assurance
ï¶ Empathy
ï¶ OthersâŠ
ï¶ Competence
ï¶ Credibility
ï¶ Security
ï¶ Access
Gaps
ï¶ Consumer expectation and
management perception
ï¶ Management perception
and service quality
specification
ï¶ Service quality specification
and service delivery
ï¶ Service delivery and
external communication
ï¶ Expected service and
experienced service
#716/04/2013 Prague, Czech Republic
8. #8
Monitoring tool
(execution environment)
Continuous
assurance
Analytics Prediction
Quality
Model
Customer
Profile
Cloud Service
Profile
Mapping &
configuration
Type of
operation
Dashboard
-Abstraction+
âŠ
Domain
knowledge
High-level
tools
Built-ins
services
+Executable-
âŠâŠ
⊠âŠ
16/04/2013 Prague, Czech Republic
Overview of
a QoS Management Architecture
9. State-of-the-Art
ï¶ Cloud Management Application Platforms
ï¶ QoS Models
ï¶ Monitoring tools and techniques
ï¶ Execution environments (Big Data analytics)
#9
âŠ
* RodrĂguez, J. M. A; Kourtesis, D.; Paraskakis, I. Semantic- based QoS management in Cloud Systems: Current Status and Future Challenges. Future Generation
Computer Systems, Special Issue on Semantic Technologies and Linked Data over Grid and Cloud Architectures. IF: 1.978 (2012). (Under review).
âŠ
16/04/2013 Prague, Czech Republic
14. Approach
#14
âą Concepts & relationships
âą Dimensions, indicators and metrics
âą Service and Customer profile
âą Reuse of existing vocabularies and standards
Abstract Model of Qos Management
âą Standard, common and shared data model
âą data integration through semantic technologies
âą Configuration
âą Dashboard
âą Qualify Functions deployment (aggregation operators)
Mapping and High-level tools
âą Monitoring tool
âą Continuous queries
âą Connection to data sources
âą ~Google analytics or Google Trends for QoS in cloud systems
Execution
16/04/2013 Prague, Czech Republic
15. 16/04/2013 Prague, Czech Republic #15
Partial Data model
Overview
*Reuse of existing models and standards are not included.
16. 16/04/2013 Prague, Czech Republic #16
I still need helpâŠ
I have a mobile application that needs a Geocoding service and
the response time must be in milliseconds.
âą More than 54 geocoding APIs
â How can I select the most suitable service?
â How can I compare different providers?
â How can I track the quality (response time) of the selected
service?
â âŠ
http://blog.programmableweb.com/2012/06/21/7-free-geocoding-apis-google-bing-yahoo-and-mapquest/
19. Key points
ï¶ Represent providers and my own QoS features in a
common, shared and standard way
ï¶ to be able to consume and make comparisons (information and data):
ï¶ E.g. compare metrics with different units, seconds and milliseconds
ï¶ Map providers API information to the QoS model
ï¶ Connectivity parameters
ï¶ Data
ï¶ Deploy the quality function and Track the services with the
monitoring tool
ï¶ Select âthe bestâ according to my target profile
#1916/04/2013 Prague, Czech Republic
21. * A toy example of monitoring the
use of words in Tweeter
#21
Storm
Trident
Real-
time
views
Batch
views
Storm
Trident
Algorithms Sync
Registered Queries
(Quality Functions)
Results
Monitoring tool
16/04/2013 Prague, Czech Republic
22. Benefits
#22
âą Integrated and Unified view of QoS
âą Extensibility
Abstract Model Qos Management
âą Standard, common and shared data model (maybe semantically-based )
âą (Semi)-Automatic deployment of Quality Functions
âą Expressivity and Analytics
Mapping and High-level tools
âą Real time capabilities
âą Big Data processing
âą Flexibility & scalability
Execution
16/04/2013 Prague, Czech Republic
24. Situated QoS
#24
⊠can a broker take advantage of the QoS
management?
16/04/2013 Prague, Czech Republic
25. Research Questions
ï¶ Which are the concepts and relationships to take into account
in QoS management?
ï¶ subjective and objective
ï¶ Which services must be provided to exploit domain
knowledge and which algorithms are necessary to afford
those services?
ï¶ How can we deal with the processing of heterogeneous data
streams (Big Data) in real-time?
ï¶ How can we find services according to customer profile
(matchmaking)?
ï¶ How can we exploit the historical information and feedback
the domain knowledge?
#2516/04/2013 Prague, Czech Republic
26. Next Steps
1. Design and deploy a complete example (iteratively)
1. Design a simple model covering some QoS features
2. Map the model and QoS features to 1 service and n providers
3. Deploy (semi-automatically) the quality function in the monitoring tool
4. Improve the monitoring tool
5. Check results
2. Go in-depth in the concept of âQualityâ and âMeasured
serviceâ
3. Look for synergies
4. Design of experiments and writing
1. Can I easily extend the QoS model? (extensibility)
2. Can I design and deploy quality analytic functions more fast? (expressivity)
3. Can I meet (first) the âcustomerâ requirements? (flexibility & scalability)
#2616/04/2013 Prague, Czech Republic
27. ï¶ Publications
ï¶ RodrĂguez, J. M. A; Kourtesis, D.; Paraskakis, I. Semantic- based QoS management in
Cloud Systems: Current Status and Future Challenges. Future Generation Computer
Systems, Special Issue on Semantic Technologies and Linked Data over Grid and Cloud
Architectures. IF: 1.978 (2012). (Under review).
ï¶ Others derivate of previous works (SCP and CHB journals)
ï¶ Talks
ï¶ Seminar at SEERC on the topic: âTowards a Pan â European E-Procurement Platform to
aggregate, publish, and search public procurement notices powered by Linked Open
Data: The Moldeas Approachâ. 22 February 2013.
ï¶ PC member and reviewer
ï¶ PC member DATAWEB (PCI 2013), ETAS 2013, ICOHT 2013 and DMoLD workshop
ï¶ Reviewer of JCR Journals: FGCS, ESWA and Current Topics in Medicinal Chemistry
ï¶ Technical Development Editor in Manning Co.
ï¶ Member of the Advisory Board in two books of IGI-Global.
ï¶ Training
ï¶ Seminar on OpenTosca
ï¶ Prototypes
ï¶ An early prototype of a real-time platform for dealing with data streams and execute simple rules is
now available (documentation and source code).
#27
Activities
16/04/2013 Prague, Czech Republic
29. ï¶ M. Maiya, S. Dasari, R. Yadav, S. Shivaprasad, and D.S.
Milojicic, "Quantifying Manageability of Cloud
Platforms", ;in Proc. IEEE CLOUD, 2012, pp.993-995.
ï¶ âA Runtime Quality Measurement Framework for Cloud
Database Service Systemsâ, 8th Int. Conf. on the Quality of
Information and Communications Technology //
Lisbon, Portugal, 2012
ï¶ N. Marz and J. Warren, âBig Data Principles and best practices
of scalable realtime data systemsâ, Manning
Publications, 2013.
#29
Main References
16/04/2013 Prague, Czech Republic