=Paper= {{Paper |id=None |storemode=property |title=Towards Integrating TOSCA and ITIL |pdfUrl=https://ceur-ws.org/Vol-1029/paper6.pdf |volume=Vol-1029 |dblpUrl=https://dblp.org/rec/conf/zeus/DemontBKLW13 }} ==Towards Integrating TOSCA and ITIL== https://ceur-ws.org/Vol-1029/paper6.pdf
          Towards Integrating TOSCA and ITIL

Christoph Demont2 , Uwe Breitenbücher1 , Oliver Kopp1 , Frank Leymann1 , and
                           Johannes Wettinger1
1
    Institute of Architecture of Application Systems, University of Stuttgart, Germany
         {breitenbuecher,kopp,leymann,wettinger}@iaas.uni-stuttgart.de
    2
      Informations- und Kommunikationszentrum, University of Stuttgart, Germany
                       christoph.demont@izus.uni-stuttgart.de



        Abstract The integration of low level management functionalities pro-
        vided by TOSCA and high level processes as defined by ITIL may provide
        significant improvement opportunities to the application provider as on
        both levels workflow technology can be employed. In this paper, we
        present Key Performance Indicator Analysis Plans as first idea how both
        approaches can be integrated.


1     Introduction and Fundamentals
Topology and Orchestration Specification for Cloud Applications (TOSCA) [3]
supports automating Cloud application management by providing a formal
method to model the structure of applications as topology and their management
functionalities through so called management plans. These plans provide man-
agement functionality such as instantiating a service, backup data or scaling out
applications. A major issue in this field is the question how to embed these plans
in the overall management strategy of the application provider. The correct time
to execute a management plan often depends on certain situations which may be
expressed as Key Performance Indicator (KPI) [2] values. In this paper we propose
an idea showing how to integrate TOSCA with the IT Infrastructure Library
(ITIL) [1] to manage the execution of plans in a well-defined but flexible manner
by introducing so called KPI Analysis Plans (KPI-APs). Our idea supports using
workflow technology on both layers to achieve robust and reliable management.
First, we explain TOSCA and ITIL and present our integration idea afterwards
before we conclude and give an outlook.
    TOSCA is an emerging standard supported by a number of prominent com-
panies in the industry such as IBM, SAP, and Hewlett-Packard. It enables the
specification of portable and holistic service models that can be used to automati-
cally instantiate concrete services in the Cloud. A service model basically specifies
a service topology (graph) consisting of nodes and relationships between nodes
to define the structure of a Cloud service. A node is any kind of component that
can be deployed into the Cloud, e.g., a virtual machine or a software component
hosted on a virtual machine. First, nodes and relationships need to be specified
by defining node types and relationship types. These type definitions include
abstract operation definitions and concrete implementation artifacts that are
attached to particular operations. As an example, a “MySQL database” node type
may own an “install” and a “configure” operation to set up a MySQL database
instance. The operations may be implemented by a set of scripts that are part of
the service model as implementation artifacts. Second, the actual service topol-
ogy is defined by connecting nodes derived from node types using relationships
derived from relationship types. An example for a very simple service topology is:
“MySQL database [hosted on] Ubuntu Linux [hosted on] virtual machine.” The
complete service model including all its parts is contained in a Cloud service
archive (CSAR) [3]. Complex composite applications can be created by combining
several CSARs. The person who is in charge of creating and maintaining the
service model may create and add management plans (workflows) to the service
model in order to define any kind of management activity. Examples for such
an activity are service deployment, database backup, or updating an application
component. Typically, these plans are defined using a workflow language such as
Business Process Model and Notation (BPMN). Plans interact with operations
defined by node types and relationship types to perform actions on a particular
service instance.
    ITIL is a widely adopted approach for IT Service Management. It provides an
accepted practical framework to identify, plan, deliver and support IT services
to the business [1]. All necessary processes, the structural organization and the
tools to be used are described. One of ITILs objectives is to gain cost benefits
for the applying organization. The framework also includes recommendations
for the definition of Key Performance Indicators (KPI) which are necessary for
measuring performance and condition of IT services. One requirement in cloud
computing is the utilization of ITIL for defining processes and to derive measures
to quantify economic benefits and impact. For monitoring and controlling IT
organization, KPIs are interpreted and measures can be taken to improve process
performance continuously. On operative level adequate tasks are derived from
improvement measures, which need to be executed at the time defined to finally
improve the IT services in a continuous improvement process [4]. approved within
a comparison of operating costs that ITIL based application management in
cloud computing supports meeting QoS requirements, safes costs and leads to
an overall better service quality. Therefore, we utilize ITIL as a management
framework to optimize Cloud application management automated by TOSCA.


2    Integration Idea

To control and manage an IT organization close to the optimum with ITIL
the organization should operate in a workflow-based process oriented way – in
organizational and technical respect. This enables benefitting from the workflow
technology’s properties such as fault handling, recoverability, and compensation
mechanisms on both levels. KPIs support the management in its decisions and are
modeled on company level as they are used to monitor and control the enterprise
performance. Both, technical and organizational KPIs are derived from defined
measures out of the processes and need to be integrated.
     KPIs are typically contracted in Service Level Agreements (SLA), are con-
tinuously renegotiated in service level management processes, and adapted to
changing business requirements. To meet these SLA requirements, the specific
and adequate measures in TOSCA need to be defined and recorded and to be
submitted to the business processes. TOSCA already works in a process oriented
manner. The plans are ideal connecting points to and from ITIL based business
processes. To connect TOSCA systematically with ITIL based business processes,
the measures coming from TOSCA need to be integrated into KPIs defined in
ITIL. This problem is tackled in this paper. Of course, vice versa mechanisms for
triggering plans and dynamic plan generation or adaption of existing plans need
to be provided in future work.
     Thus, the integration of TOSCA and ITIL consists of two cyclic steps influenc-
ing each other directly: Analysis and Monitoring of Key Performance Indicators
and triggering TOSCA management plans based on these results to react to the
measured KPIs. The analysis of an application’s KPIs is typically a complex
challenge. Although there are monitoring frameworks and tools, the integration
of different KPI measurements of different application components is difficult
and mainly depends on the overall characteristic of an individual application: For
some applications a certain combination of KPIs may be appropriate while other
applications would lose their key success factors. Thus, there are two challenges:
(i) Integrating different KPIs measurements of different application components
in a (ii) customizable but well-defined manner. We argue, that a generic approach
is not suitable as especially the ability to customize the integration based on
individual requirements is of vital importance as stated above. Thus, to tackle
this problem, we introduce an analysis method which is tightly coupled to an
individual application but benefits from standardized and reusable artifacts:
Key Performance Indicator Analysis Plans (KPI-AP). These plans are TOSCA
management plans implemented by the application developer of the application
itself following well-defined KPI metrics. They are responsible for measuring
high level KPIs of the overall application by orchestrating individual low level
KPI measurement operations provided by the application’s components and pro-
vide a standardized way to integrate reusable artifacts and enable portable KPI
measurement as they are contained in CSARs. This allows different application
providers to embed the plans in their overall ITIL management processes as their
functionality and KPI results are well-defined.


3    Conclusion and Outlook

In this paper we presented a first idea showing how to integrate TOSCA and
ITIL by using Key Performance Indicator Analysis Plans. These plans provide
a portable and self-contained way for integrating KPI measurements into the
application providers overall management. In future work, we want to analyze how
to embed requesting TOSCA management fully automated into ITIL processes
and how the presented KPI-APs may be employed to achieve this.
References
1. APM Group Ltd: Official ITIL Website (2013), http://www.itil-officialsite.
   com/
2. Gabler Wirtschaftslexikon: Key Performance Indicator (KPI) (2013), http://
   wirtschaftslexikon.gabler.de/Definition/key-performance-indicator-kpi.
   html
3. OASIS: Topology and Orchestration Specification for Cloud Applications Version
   1.0. Working Draft 14 (November 2012), http://www.tosca-open.org
4. Rajan, S.S.: Cloud Computing Application Management and Lean ITIL. Cloud
   Computing Journal (October 2011)

   All links were last followed on April 2, 2013