Integrated Lifecycle of IT Services in
a Cloud Environment
Karuna P Joshi
Ebiquity Research Group,
Computer Science and Electrical Engineering Dept, UMBC
http://ebiquity.umbc.edu/; http://mc2.umbc.edu
kjoshi1@umbc.edu
Introduction: Software Delivery Models
In-house approach: Traditionally, software applications
development and maintenance has been done within
the company with it’s own employees.
 Outsourced or Off shored approach: Software
applications purchased by companies. Development
and maintenance is contracted to consulting/staffing
companies who could be based anywhere in the
world.
 Services on the Cloud approach: IT services
purchased from software manufacturers who could
be a single entity or a mix. Also termed virtualized
software to imply the software development and
maintenance is transparent to the consumer.

IT Services

Information Technology (IT) increasingly being
viewed as a “Service” offered to an organization
 Software, Customizations, Support, Hosting and
Maintenance purchased as a “Service” from providers.

The X as a Service alphabet soup
◦
◦
◦
◦

Infrastructure as a Service (IaaS)
Platform as a Service (PaaS)
Database as a Service (DaaS)
Software as a Service (SaaS) etc…
Services are virtualized -- shared across multiple
provisioning.
Virtualized IT Services on the Cloud
Composed of one or more components
developed by one or more providers and
delivered as a single service to the consumer.
 Service dynamically composed

◦ Service structure/components not pre-determined
◦ Based on consumer’s requirements. (On demand
composition)


Delivered virtually via the Cloud / grid.
Consumer specifies both functional and non
functional requirements like quality, security,
compliance etc.
Key Open Issues

Composition and orchestration of services currently
requires a programmer with knowledge of
◦ WSDL (Web Services Definition Language) or similar languages to
describe the services.
◦ BPEL (Business Process Execution Language) to specify the
interactions between the services.
◦ Users who are not tech-savvy cannot use this.

Currently there is no negotiation capability for service
composition in cloud environment
◦ Cannot define variable constraints or negotiate on cost, SLAs
etc.

No automation of the current approach to discover
services like RFPs, etc.
Our Approach

We propose a policy-based approach to automate the
service lifecycle on the Cloud.

Use Semantic web languages/technologies like OWL (Web
Ontology Language), RDF (Resource Description
Framework) and associated reasoning infrastructure such as
AIR, Jess, etc.

Extend the Service Oriented Architecture (SOA) to
factor in expressive policies, negotiation and rules-based
reasoning to make it more suitable for Cloud
Environment.
Policy for Services on the Cloud
Enterprise Policies
Service
Negotiation
• Business Processes
• Budgetary rules
• HR policies
• Infrastructure / Facilities
• Security
• Compliance etc.
Policies
CONSUMER
• Service Cost
• Service Agreements
• Service Obligations
• Security
• Compliance, Taxes etc.
• Request for
Service
• Constraints
Request
For service
Enterprise IT
Service Environment
• Services
• Packaged applications
• User devices (e.g. PCs,
Laptops, PDAs etc.)
• Servers
• Network Infrastructure
Service Policies
Service Level
Agreement
Provider Service
Environment
Service
Presentation
service
• SaaS
• DaaS
• PaaS
• IaaS
SERVICE CLOUD
Policies
service
• Service components
• Dependent Services
• Service Agents
• Development platform
• Staging platform
PROVIDER
Methodology for Service Lifecycle

Our proposed methodology divides Service
Lifecycle into Five Phases
◦
◦
◦
◦
◦
Requirements
Discovery
Negotiation
Composition
Consumption
This Methodology is applicable on private or public
cloud.
● Have proposed metrics to track success of each
phase.
● Have developed initial version of ontologies for the
five phases.
●
Phases of IT Services Lifecycle
Service
Requirements
Service specified
Service Discovery
Provider(s) identified
Service Negotiation
Contract signed
New Service needed
Service Consumption
CONSUMER
Service Composition
Service delivered
SERVICE CLOUD
Services Requirements phase
‘Request for
Service’ Monitor
Policy Manager
policy
consumer
service
•Policy Type
•Hard constraints
•Soft constraints
•Dependents
•Parents
Services Manager
•Name
•Provider
•Type
•Attributes
•Location
•Dependencies
Consumer
•Service
•Requirements
policy
Services
Procurer
•Service specs
•Policies with
constraints
• RFS generator
specs
Provider 1
‘Request for
Service’ Monitor
Issue
Request
for
Service
(RFS)
•Service
•Requirements
Provider 2
‘Request for
Service’ Monitor
•Service
•Requirements
Provider 3
‘Request for
Service’ Monitor
•Service
•Requirements
Provider 4
Service Cloud
Service Requirements
 Requirements
for a service will include
◦ Functional specifications (tasks to be automated)
◦ Technical specifications




Software applications and constraints
Hardware Policy – Mobile device, PC, Server, Multicore etc.
Operating System Policy – single OS, multiple OS
Language Support Policy
◦
◦
◦
◦
Agent experience
Agent skill level
Geographical location constraints
Nationality/Work permit constraints
◦ Human Agent specifications
Service Requirements (contd…)
 Requirements
◦ Security Policy
for a service will include
◦ Role
◦ Permissions
◦ Data Quality Policy
◦ Low quality data may be acceptable to consumer if it provides
cost saving
◦ Service Compliance Policy
 Standards adhered
 Certifications needed
 Government regulations adhered.
Example of Requirements Policy
Data Quality policy - // allows services of low data quality to be accepted if the service is free of cost.
// Using AIR Policy Language
@forAll :SERVICE_QUALITY, :COST.
:data_quality_policy a air:Policy;
rdfs:label "Data Quality policy";
air:rule :data-quality-rule.
:data-quality-rule a air:Belief-rule;
rdfs:label "data quality rule";
air:pattern {
:SERVICE_QUALITY quality_level :LOW.
:COST service_cost :0.
};
air:assert {:SERVICE_QUALITY air:compliant-with :data_quality_policy.};
air:alt { air:assert {:SERVICE_QUALITY air:NOTcompliantwith :data_quality_policy.}}.
High Level Ontology for Service Requirement Phase
Class : Request for Service
Class:
Specification
Class:
Service
Compliance
Name
Description
is part of
is part of
is part of
is part of
Class:
Data
Quality
policy
Class :
Functional
Specs
Task
Description
is part of
Class: Human
Agent Specs
Experience
Availability
Nationality
Other
Constraints
Class: Domain
Domain
Characteristics
Is part of
Is part of
Service Domain
Exp_Svc_Begin_Date
Exp_Svc_End_Date
RFS_Respond_by_dt
 Cost_constraint
Must
include
is part of
is part of
Class
Technical
Specs
Class:
Security
policy
Is part of
is part of
Class:
Operating
System
Details
is sues
Class:
Consumer
Consumer-ID
Description
is part of
Class:
Hardware
Details
Class:
Compatible
Application
Details
Service Discovery

Services search/discovery engine
◦ Used to search available services that match the
specifications
◦ Identify functional gaps that exist in services discovered

Service Certification
◦ A central registry, like UDDI, will certify a service provided.
◦ Consumers can refer to this certification when selecting a
service.
Service Discovery phase
Service Verification /
Certification
Service
Requirements
Service Discovery Engine
List of service providers
with advertised service,
service levels and cost
Service
Negotiation
CONSUMER
SERVICE CLOUD
Policy in Service Discovery phase
RFS Responder
Policy Manager
consumer
Services
Manager
service
•Name
•Provider
•Type
•Attributes
•Location
•Dependencies
Consumer
RFS
•RFS
negotiation
specs
Service
available
Cannot
provide
service
Cost
Availability
Functionality
Requirements met
Provider 1
RFS Responder
• Availability
Provider 2
Service
available
Service
Provider
Validation
•Service
•Provider
•Validation
RFS Responder
• Cost
• Availability
Functionality,
Requirements met
Service
Composer
•
•
•
•
Components
Constraints
Functionality
Requirements met
Provider 3
+
policy
•Policy Type
Services
•Hard
Procurer
constraints
•Soft constraints policy •Service
•Dependents
•Policies
•Parents
•Generate
•
•
•
•
Provider 4
Service Cloud
High Level Ontology for Service Discovery Phase
Class : Request for Service
Service Domain
Exp_Svc_Begin_Date
Exp_Svc_End_Date
RFS_Respond_by_dt
 Cost_constraint
returns
list to
Class: Consumer
Consumer-ID
Description
Class: Service
Certification
used in
Class:
Service Search
Engine
Verifies
Provider
creates
Provider
Service Name
Service ranking
adds to
Class : Provider List
Class: Provider
Provider
Service details
Service availability
Service Cost
Service list
Description
Service Negotiation
Discussion and agreement that the
Service provider and consumer have
regarding the Service.
 Service Level Agreements (SLA) finalized
between consumer and provider
 Quality of Service (QoS) decided
between primary provider and
component providers.

Service Negotiation phase
Service
Discovery
Service Level Agreement
(SLA) between consumer
and primary service provider
Service Quality of Service (QoS)
between primary service
providers and dependent services
Service Composition,
Orchestration
CONSUMER
SERVICE CLOUD
Policy in Service Negotiation phase
Iterative
process
Policy
Manager
policy
•Policy Type
•Hard
constraints
•Soft constraints
•Dependents
•Parents
consume
r
Services
Manager
service
•Name
•Provider
•Type
•Attributes
•Location
•Dependencies
policy
Services
Procurer
•Service
•Policies
•Generate
RFS
specs
•RFS
negotiation
Revised soft
constraints,
SLA
Revised cost,
configuration
Iterative
process
Revised soft
constraints,
SLA
Revised cost,
configuration
Final cost,
SLA,
configuration
Service
Negotiator
•Cost
•Availability
•Service
functionality
•Requirements
met
Service
Negotiator
•Cost
•Availability
•Service
functionality
•Requirements
met
Provider 1
Policy
Manager
policy •Policy Type
•Hard constraints
•Soft constraints
•Dependents
•Parents
policy
Policy
Manager
•Policy Type
•Hard constraints
•Soft constraints
•Dependents
•Parents
Provider 3 Provider 4
Consumer
Service
Cloud
Provider
High Level Ontology for Service Negotiation Phase
Class : Request for Service
Service Domain
Exp_Svc_Begin_Date
Exp_Svc_End_Date
RFS_Respond_by_dt
 Cost_constraint
Consumer-ID
Description
initiates
Class:
used in
Class:
Contract
Negotiation
Class:
Consumer
subClass of
subClass of
used in
Class : Provider List
Provider
Service details
Service availability
Service Cost
Consumer
Negotiation
results
in
participates
Class:
Service Level
Agreement
Class :
is part
of
SLA Name
Description
SLA Metrics
Penalty
subClass of
Class:
Contract
Class: Provider
Service list
Description
Class :
Quality of
Service (QOS)
initiates
Class:
Provider
Negotiation
results
in
QOS Name
Description
QOS Metrics
Penalty
Service
Contract
subClass of
is part
of
Class :
Dependent
Service
SubContract
Service Composition Phase
One or more services provided by one or more
providers are combined and delivered as a single
Service.
● Service orchestration determines the sequence of
execution of the Service components.
● Service Coupling measure determines how
dependent main service is on other services or
resources for its delivery
● Service Cohesion measures the extent to which
related aspects of a requirement are kept together in
the same service, and unrelated aspects are kept out.
●
Service Composition phase
Service
Negotiation
Service
composed
Dependant
services
Service
Consumption
CONSUMER
SERVICE CLOUD
High Level Ontology for Service Composition Phase
Class:
Specification
Name
Description
Determines
Class: Provider
Service list
Description
Executes
Defines
metrics
for
Class : OWL-S
– Composite
Process
Is part
of
Class: Service
Level Agreement
SLA Name
Description
SLA Metrics
Penalty
Class : Quality of
Service (QOS)
Class:
Dependent
Service
Defines
metrics
for
QOS Name
Description
QOS Metrics
Penalty
Service Consumption Phase
 Composed
Service is consumed and
monitored in this phase.
 Key measures like Service Performance
and reliability are monitored using
automated tools.
 Phase includes Service Delivery, Service
payment.
 Customer Satisfaction is tracked in this
phase.
Service Consumption phase
Service
Composition
Service
Requirements
Service Monitored
Service delivered – one time
or periodically as needed
Service consumed
Service
payment
CONSUMER
SERVICE CLOUD
High Level Ontology for Service Consumption Phase
Class: Provider
Service list
Description
Provides
Class:
Consumer
Consumer-ID
Description
is consumed
by
Class:
Dependent
Components
Class :
Response
Time
Class :
Service
Contract
Class : OWL-S
– Composite
Process
Class :
Performance
Determine quality for
Class :
Software
components
Class:
Reliability
Class :
Latency
Class :
Throughput
refers to
Class:
Security
Policy
Class:
Resolution Time
Part of
Class : Service
Level Agreement
Task
Description
Class : Tier
Class: Human
Component
Class:
Presentation
Class:
Assurance
subClass
of
subClass
of
1 agent
Class : Tier
2/ external
agent
Service Life cycle: detailed
“Request for Service”
Identify functional and
technical specifications
Determine domain,
data type and it’s
acceptable quality
levels
Service Discovery
Engine
Service
Certification
List of service providers with advertised
service, service levels and cost
Service Level Agreement (SLA) between
consumer and primary service provider
Service Monitoring
Service consumed
CONSUMER
Service
payment
Quality of Service (QoS) contracts between
primary service providers and dependent services
Service composed
Service packaged,
delivered – one time or
periodically as needed
SERVICE CLOUD
Dependant
services
Ongoing Work
Currently developing policies for each phase of
the service lifecycle using policy languages like
AIR policy language developed by MIT.
 We continue to refine the IT Service ontology.
Ontology in OWL for the entire Lifecycle is
available at

http://www.cs.umbc.edu/~kjoshi1/IT_Service_Ontology.owl

Developing an integrated quality framework for
virtualized services delivered via the cloud.
Questions ?

Relevant Publications on the Ebiquity
website (http://ebiquity.umbc.edu)