On Analyzing and Specifying Concerns for Data as a Service

On Analyzing and Specifying Concerns
for Data as a Service
Hong-Linh Truong and Schahram Dustdar

Distributed Systems Group
Vienna University of Technology

truong@infosys.tuwien.ac.at
http://www.infosys.tuwien.ac.at/Staff/truong/
Acknowledgment: Marco Comerio, G.R. Gangadharan, Roman
Khazankin, Reinhard Pichler, Andrea Maurino, Vadim Savenkov,
ttp://www.VitaLab.tuwien.ac.at

1

Outline

 Background and motivation
 DaaS concerns
 Specifying DaaS concerns
 Linking DaaS concerns to services
 Current prototype
 Some studies of DaaS concerns in current
service descriptions
 Conclusion and future work

2
APSCC 2009, Dec 8th , 2009, Singapore

Background

 Web services technologies, the SaaS model and
the cloud computing model foster the concept of
data/information as a service (DaaS)‫‏‬
 No precise definition but DaaSs
 Provide data capabilities rather than provide
computation on data or data based on computation
 Providing DaaS is an increasing trend
 In both business and e-science environments
 Bio data, weather data, company balance sheets, etc., via
Web services
 Academic research and industrial relevant research
topics
3

Background - our view on DaaS

 Read-only DaaS versus CRUD DaaS
 Service APIs versus Data
 Service APIs are used to CRUD data
 They are not the same wrt concerns

http://www.strikeiron.com/ 4 http://infochimps.org/

Motivation

 Data-specific concerns need for
 Selecting data services based on provided data and
service contracts
 Evaluating the compatibility of service contracts in
data composition
 Supporting quality-aware data composition from
multiple data services
 Data-specific concerns combined with service
APIs specific concerns
 Not just QoS based service selection

5

Motivation (cont.)

 DaaSs are currently considered like any other Web
services
 WSDL/WADL description + QoS + pricing information
(mostly in HTML form)
 But concerns on data are different from that on service
APIs
 Where are the data-relevant concerns in service
descriptions?
 E.g., data quality, usage permission, and data
ownerships
 How data-relevant concerns can be combined with
service-relevant concerns?
6

Existing Work
 QoS description and QoS-based Web services selection are well
researched
 Googling "QoS-based‫‏‬Web‫‏‬services‫‏‬selection„‫000.02‏~‏‬
 Data Quality is well-known in database community
 E.g., see ACM Computing Survey 41(3):2009 on data qualities
done by Batini et al.
 (Service) Licensing is currently being studied for SaaS
 Several licenses for data are introduced but in human-readable form
only
 E.g., Talis community license, the Open Knowledge Foundation
Wiki, the Open Database License
 Intensive discussion on laws and regulations on cloud computing
 E.g.,‫‏‬see‫‏‬Davide‫‏‬Maria‫‏‬Parrilli‘s‫‏‬work
 Data Governance: e.g., see the IBM data governance maturity model
7

Issues and Approach
 Issues
 DaaS concerns include QoS, DQ, service licensing, data licensing,
data governance, etc.
 There is a lack of techniques for the publishing, discovery,
selection and evaluation of data concerns
 There is a lack of techniques for integrating concerns for DaaSs
 Data concerns and Service APIs concerns
 This talk focuses on publishing information that characterizes DaaSs
What are main DaaS concerns (non-functional parameters) and how
to specify them and provide them for the data service selection and
contract compatibility?
Some empirical studies on existing DaaS descriptions
We are not talking about how to evaluate concerns and monitor them

8

The Importance of Concerns in
Data Consumer‘s View
Concerns Read-only DaaS CRUD Daas

Data Quality Important factor for the selection of Expected some support to
DaaS. For example, the accurary control the quality of the data in
and compleness of the data, case the data is offered to other
whether the data is up-to-date consumers
Data source Important factor for the
trustworthiness of the DaaS.
Data & Service Important factor, in particular, price, Important factor, in paricular,
Usage data and service APIs licensing, law price, service APIs licensing, and
enforcement, and IPRs law enforcement
Data Governance Important factor, for example,
the security and privacy
compliance, data distribution,
and auditing
QoS Important factor, in particular Important factor, in particular,
availability and response time availability, response time,
depability, and security
Service Context Useful factor, such as classification Important factor, e.g. location
and service type (REST, SOAP), (for regulation compliance)‫‏‬and
APSCC 2009, Dec 8th ,
location
2009, Singapore
9 versioning

Conceptual Model for DaaS
Concerns and Contracts

10

Capability Concerns

 Data Quality capabilities
 Based on well-established research on data quality
 Timelineness, uptodate, free-of-error, cleaning,
consistency, completeness, domain-specific
metrics, etc.
 We mainly support the specification of DQ metrics for
the whole DaaS but possible to extend to the service
operation level
 Data Security/Privacy capabilities
 Data protection within DaaS, e.g. encryption, sensitive
data filtering, and data privacy
 Many terms are based on the W3C P3P
11

Capability Concerns (cont.)

 Auditing capabilities
 Logging, reporting (e.g., daily, weekly, and monthly),
and warning
 Support system maintenance, SLA monitoring, billing,
and taxation
 Data lifecycle
 Backup/recovery, distribution (e.g., a service is in
Europe but data is stored in US), and disposition
 Support system maintenance but also regulation on
data

12


 QoS capabilities are applied to service APIs
 Based on well-researched QoS for Web services
 Performance capabilities
 e.g., latency, response time and throughput
 Dependability capabilities,
 e.g., availability, reliability, accessibility, security
 Business
 Pricing model (flat rate, pay-per-use, with/without
transaction conditions)‫‏‬and Price
 Service credit for reward or compensation
 e.g. Amazon service credits
13


 Data and service license
 Usage permission: for data (distribution, transfer,
personal use, etc.) and for service APIs (adaptation,
composition, derivation, etc.)‫‏‬
 We utilize some terms from ODRL/ODRL-S
 Copyrights
 Liability: e.g., who is reponsible for the loss due to a
network disruption?
 Law enforcement (e.g., US or European court)‫‏‬
 Domain specific IRPs

14

Data Source Concerns

 A DaaS may utilize data from many sources.
 Similar DaaSs may utilize data from the same source
 Data source properties
 Name: e.g. ddfFlus or DataFlux or Mr A
 Size
 Timespan: the duration of collected data, e.g., more
than 4 years in the eBay Data License
 Update Frequency: how offen the data is updated
 Etc.

15

Service Context Concerns

 Location:
 Selecting a DaaS in Amazon US Zone or European
Zone?
 Service Type: REST or SOAP?
 E.g., mobile client daas
 Level of Service
 Service Classification
 Based on UNSPSC Code Classification Services
 Data Classification
 Service/data versioning

16

XML Diagram for DaaS Specification

17

XML Diagram for the DaaS
Capability Specification

18

From Capability/Context to
Service Contract
Non-functional parameters (NFPs) to Service Contracts

Define and
Search NFPs
negotiate contract Contracts
of DaaSs
terms

DaaS Capabilities,
Context, Data Consumer-specific
19 concerns
Source

 A service contract includes a set of generic,
data-specific and service-specific conditions
established based on concerns

Populating DaaS Concerns
The role of stakeholders in the most trivial view
evaluate, specify,
publish and manage Consumer
DaaS Provider
specify, select,
monitor, evaluate
DaaS
Concerns

monitor and
evaluate Third-party service

 We address the specification, publishing and
management of DaaS concerns
 To support the selection of DaaSs
 Monitoring and evaluation are currently open
20

Implementation
 Concern specifications
 Possible solutions: XML, RDF, and OWL
 Our implementation is based on XML/RDF
 Easy to reuse vocabularies defined in other
standards
 Link to external domain-specific models of
concerns using URIs
 Publishing and linking concerns to services
 Possible solutions: annotating WSDL, SAWSDL, and
external management services
 We use our SEMF model. Concerns are managed
via services supporting the evolutionary management
21

Example of linking concerns
with other type of data
Based on SEMF (Service Evolution Management Framework) [SEAA 08]
<t i t l e>Co r t e r aCr e d i t P u l s e S e r v i c e</ t i t l e>
<e n t r y>
<t i t l e>I n t e r f a c e</ t i t l e>
<summary>WSDL I n t e r f a c e </ summary>
<c a t e g o r y l a b e l ="Web Service Description" scheme="http://www.dmoz.org/Computers/
Programming/Internet/Service-Oriented_Architecture/Web_Services/WSDL“
t e rm="Interface" />
<c o n t e n t t y p e ="application/wsdl+xml" s r c ="http://ws.strikeiron.com/
CorteraCreditPulse2?WSDL" />
</ e n t r y>
<e n t r y>
<t i t l e>DaaS Conc e rns</ t i t l e>
<summary>Data Conc e rns</ summary>
<c a t e g o r y l a b e l ="Data Concerns" t e rm="DaaSConcern" />
<c o n t e n t t y p e ="application/xml" s r c ="http://www.infosys.tuwien.ac.at/prototyp/SOD1/
dataconcerns/samples/CorteraCreditPulseConcerns.xml" />
</ e n t r y>

22
APSCC 2009, Dec 7-11 2009, Singapore

Support DaaS Concerns Selection
Data SECO2
Consumer

DeXIN

Service Information
Management
Service

SEMF-based External
information, including sources
concerns

 DeXIN: Distributed XQuery over Heterogeneous Data Sources [ICEIS09,
ICWE09]
 SECO2 : Service Contract Compatibility [ICSOC09]
23

Some Studies

 We are not aware of any provider that publishes
DaaS‘s‫‏‬concerns‫‏‬in‫‏‬a‫‏‬well-defined form
 Only HTML
 Our studies examines the description of DaaSs
 Enterprising computing
 StrikeIron, Xignite, serviceobjects.NET, WebserviceX,
XWebServices, AERS, Amazon
 E-science
 GBIF (Global Biodiversity Information Facility), EBI
(European Bioinformatics Institute) Web Services,
EMBRACE Service Registry, and BioCatalogue
24

Service Classification

 StrikeIron
Web
services

 Xignite
Web
services

25

Service Classification
 ServiceObjects
Web Services

 WebservicesX Web services  XWebService Web services

26

0
5
10
15
20
25
30
35
Completeness

Uptodate

Correctness

Cleaning

Standard output

Privacy
based
Logging

Reporting

Warning

Backup

Response Time

Availability

Network Latency

Packet Loss

Network Security

Price Model

27
Service Credit

Usage Permission

Copyright

Liability

Law Enforcement

Domain-specific IPR

Location

Service Type

Data Classification

Data Source Name

Data Source Size
Concerns in HTML descriptions

Data Source Update Freq.
 29 services from 7 providers, most are SOAP-

Mentioned
Not mentioned/clear

Concerns of DaaSs in E-science

From the DaaS description point of view
Service Registries DQ QoS Business Licensing
Ownership Usage
permission
GBIF No No No unstructured unstructured
EBI Web Services No No No No No
EMBRACE Service No No No No No
Registry
BioCatalogue No No unstructured unstructured unstructured

28

Conclusion and Future Work
 This paper presents
 The importance of having DaaS concerns to be explicitly specified
an a study of existing concerns
 A specification and management technique for DaaS concerns
 Future work‫‏‬
 Enhance empirical studies on current concerns for DaaSs
 Apply DaaS concerns to bioinformatic and biomechanic DaaSs
 Support DaaS concern in data composition/mashup tools and
contract compatibility evaluation
 Develop a service engineering approach for DaaS concerns, and
concern monitoring and evaluation
 Need a joint effort between service engineering and data
engineering research
http://www.infosys.tuwien.ac.at/prototyp/SOD1/
29

Thanks for your attention!

Hong-Linh Truong
Distributed Systems Group
Vienna University of Technology

truong@infosys.tuwien.ac.at
http://www.infosys.tuwien.ac.at/Staff/truong/

Austria

30

On Analyzing and Specifying Concerns for Data as a Service

Recommended

More Related Content

What's hot (20)

Viewers also liked (9)

Similar to On Analyzing and Specifying Concerns for Data as a Service (20)

More from Hong-Linh Truong (20)

Recently uploaded (20)

On Analyzing and Specifying Concerns for Data as a Service