Seite 161 - Cloud Migration Version 2012 english

161 
After the Cloud service has been implemented and has assumed productive 
operation, the following topics become relevant: 
Support: The use of Cloud services primarily means a change in 
service processes. In this regard, it must be defined who will be 
responsible to document and communicate, as well as possibly scale 
up the service requirements, and how this is to be done. The 
different support possibilities will have already been clarified during 
the process of selecting a Cloud service and Cloud provider, and 
must be communicated to the users.  
Performance: Potential bottlenecks should be proactively 
communicated by the Cloud provider on the one hand, and, if 
necessary, be double checked by internal performance 
measurements. In this respect, according to the defined service level 
agreements (SLAs), the person responsible for the service is required 
to verify compliance with the performance and availability 
guarantees as well as to scale to more capacity as necessary. 
Availability: In a continuous cycle, the service provider will 
implement patches and updates. The respective information must be 
made available by the service provider and users of the service must 
synchronise it with their own update cycles. 
Information on compliance with the SLAs and parameters is generally 
delivered by the Cloud provider. The company itself is responsible for 
analysis and monitoring. The respective measures must be coordinated 
between the company and Cloud providers and should be intended to 
achieve compliance with the SLAs agreed upon.  
Identifying performance indicators 
The indicators and parameters necessary for monitoring Cloud services 
should already have been defined as "IT quantities" in the preparation 
process. They can also be identified by a careful review of the contracts with 
the Cloud provider. These parameters are defined and documented in 
service level agreements (SLAs). 
Typical metrics for Cloud service performance are: 
Availability in % of the time that the service is in operation. 
Guaranteed up‐and download bandwidth. 
Response times in the event of failure.