![](bg8.png)
form supporting the construction of DASMs for de-
ployment. They use model-based patterns to capture
abstract deployment topologies. In contrast to our
work, they focus on deployment and creating DASMs
by composing topology patterns. We apply transfor-
mation patterns to transform ASMs to DASMs.
The CHAMPS System (Keller et al., 2004) focuses
on Change Management which modifies IT systems
through so-called Requests For Change (RFC), e. g.,
installation, upgrade, or configuration requests. RFCs
are on a lower level than Management Patterns and
targeted to a single component or small group of com-
ponents while Management Patterns are intended to
affect multiple components directly at once. After re-
ceiving an RFC, the CHAMPS System assesses the
impact of the RFC on components by analyzing the de-
pendencies between the directly affected components
and their neighbours and generates a so-called Task
Graph which is afterwards used to generate an exe-
cutable Change Plan. Thus, the Desired Application
State Model is not generated by a single operation like
applying a Management Pattern but evolves by ana-
lyzing the influences of an RFC to other components
recursively. In contrast, we enforce a strict separa-
tion between change requests, which are expressed
as patterns in our work, the Desired Application State
Model, and the operational model which is represented
by planlets. This allows capturing high level expert
knowledge much more concrete in the automated exe-
cutable transformations of patterns.
The DevOps community also provides higher level
tooling such as Marionette Collective. These tools pro-
vide more convenience to manage application topolo-
gies as it turned out that large and complex topologies
are hard to manage with plain configuration manage-
ment only (Loope, 2011). The larger the topologies,
the higher the probability of mistakes when reusing ar-
tifacts. However, all these tools do not provide the high
level of abstraction provided by Management Patterns.
5 CONCLUSIONS AND
OUTLOOK
In this paper, we presented how high level and low
level management tasks can be implemented separately
in a generic and reusable manner by using Manage-
ment Patterns and Management Planlets and that they
are applicable fully automated to different applications.
To evaluate the approach, we implemented a prototype
based on TOSCA and showed that the concept also
enables application developers to influence the actual
management by implementing custom Management
Planlets. In future work, we plan to extend the sys-
tem towards applying patterns fully automated based
on certain application states and rules, e. g., to handle
peak workloads. In addition, multiple patterns should
be applicable at once.
ACKNOWLEDGEMENTS
This work was partially funded by the BMWi project
CloudCycle (01MD11023).
REFERENCES
Alexander, C. et al. (1977). A Pattern Language. Towns,
Buildings, Construction. Oxford University Press.
Arnold, W., Eilam, T., Kalantar, M., Konstantinou, A. V., and
Totok, A. A. (2007). Pattern based SOA deployment.
In ICSOC. Springer-Verlag.
Binz, T., Breiter, G., Leymann, F., and Spatzier, T. (2012).
Portable Cloud services using TOSCA. IEEE Internet
Computing, 16(03):80–85.
Breitenb
¨
ucher, U., Binz, T., Kopp, O., Leymann, F., and
Schumm, D. (2012). Vino4TOSCA: A visual notation
for application topologies based on TOSCA. In CoopIS.
Springer-Verlag.
Eilam, T., Elder, M., Konstantinou, A. V., and Snible, E. C.
(2011). Pattern-based composite application deploy-
ment. In Integrated Network Management. IEEE.
Hohpe, G. and Woolf, B. (2003). Enterprise Integration Pat-
terns: Designing, Building, and Deploying Messaging
Solutions. Addison-Wesley.
Keller, A. and Badonnel, R. (2004). Automating the pro-
visioning of application services with the BPEL4WS
workflow language. In DSOM. Springer.
Keller, A., Hellerstein, J. L., Wolf, J. L., et al. (2004). The
CHAMPS system: change management with planning
and scheduling. In NOMS. IEEE.
Loope, J. (2011). Managing Infrastructure with Puppet.
O’Reilly Media, Inc.
Maghraoui, K. E., Meghranjani, A., Eilam, T., and Konstanti-
nou, E. V. (2006). Model driven provisioning: Bridg-
ing the gap between declarative object models and
procedural provisioning tools. In ACM/IFIP/USENIX
Middleware 2006.
OASIS (2012). Topology and Orchestration Specification
for Cloud Applications Version 1.0 Committee Specifi-
cation Draft 03.
P. Cordella, L., Foggia, P., Sansone, C., and Vento, M. (2004).
A (sub)graph isomorphism algorithm for matching
large graphs. IEEE Trans. Pattern Anal. Mach. Intell.,
26(10):1367–1372.
Weld, D. S. (1994). An introduction to least commitment
planning. AI Magazine, 15(4):27–61.
CLOSER2013-3rdInternationalConferenceonCloudComputingandServicesScience
482