News‎ > ‎

Successful 3rd SmartenIT Meeting in Krakow - Scenario Descriptions

posted Jun 3, 2013, 6:50 AM by Corinna Schmitt   [ updated Jul 21, 2014, 7:43 AM ]
The partners of the SmartenIT project met in Krakow (Poland) during June 3rd - 5th, 2013.
The event was hosted by the partner AGH.

Main focus of the 3rd SmartenIT meeting was on scenarios and use-cases for SmartenIT. The partners specified in more detail the following scenarios including different use-cases (scenario leaders are mentioned in brackets):

Scenario 1: Inter-cloud communication (IRT)
The inter-cloud communication scenario focuses on the interaction between cloud operator entities (owning and managing Datacenter infrastructures) acting in a multi domain operation topology. Targeted Cloud operators to act in the scenario are small-medium size companies. The scenario is focused on a ‘Federated cloud’ model in which different cloud operators join together to allow resource sharing between the different cloud players with the final aim to deliver services with expected SLA (Service level Agreement) to their end-user.

The main topics of the selected scenario are:
  • The interaction of the formed overlay network with the underlying connectivity infrastructure offered by Internet Service Providers. SmartenIT Traffic management will provide optimized path selection to allow fair utilization of ISP resources, will achieve an improvement in QoE as perceived by the final user and will provide Energy efficiency enhancements within ISP and Cloud Operators domains.
  • Business relationships analysis among the different stakeholders relevant for the scenario (Cloud Service provider, Cloud Operator, Internet Service Provider, End-user)
  • SLA/OLA (Service Level Agreements and Operational Level agreements) impacts in the interaction among different players.
Currently the task force dedicated to the actual scenario has selected the following main use-cases:
  • Caching Service in the cloud environment
  • Bulk transfer between Cloud operators (to provide an enhanced disaster recovery service to business end-users) 
  • Static resource allocation within the Cloud Operator federation to allow sharing and offloading of resources 
  • Dynamic resource allocation within the Cloud Operators federation to mitigate peaks in resources demand/utilization as well as for energy efficiency purposes

Scenario 2: Global service mobility (TUD)

Futher information will follow soon!

Scenario 3: Exploiting online social networks information - social awareness (UniWue):
Social awareness links social signals and information such as social network structure, users' preferences and behaviours, etc. to network management mechanisms. This means that socially-aware mechanisms exploit this information in order to perform efficient network management, content placement, and traffic optimization to enhance the performance of an overlay application (e.g. video streaming, file sharing).
With respect to a SmartenIT socially-aware traffic management solution, the consortium agreed to first investigate the exploitation of social information, i.e. to assume that perfect social information is available (from a black box social information provider). After relevant mechanisms have been developed, the focus shall be moved towards how the necessary social information can be collected/harvested. This approach is in accordance with the initial SmartenIT system architecture (D3.1, Figure 35) in which the social analyzer module can process information which are provided by an overlay social network (i.e. the black box social information provider) and collected by a social monitor module.
Currently the scenario task force agreed to focus mainly on two major use cases: data offloading to WiFi and content caching/pre-fetching. A traffic management solution which will be developed in SmartenIT should exploit social information and address these use cases.

Scenario 4: Collaboration for energy efficiency (AUEB)
Scenario 4, entitled "Collaboration for energy efficiency", focuses on federation of data centres and live migration of workload and virtual machines so as to discover new effective ways to manage and optimize energy consumption. The collaboration between them will allow establishing new energy saving policies and procedures dealing with energy demand patterns of data centres and collaboration with energy producers. Apart from energy saving, also the exploitation of renewable energy sources may have high priority.
Federated data centres create a decentralized energy ecosystem where the workload can be exchanged to address energy shaping requests from the energy providers, economical demands (lower costs of energy in certain locations) or prioritizing access to stable renewable energy sources.
Advanced adaptive power consumption requires close collaboration between data centres and energy providers. There are already efforts introducing interfaces between those two sides to share the information and dynamically adjust power consumption.
The information about energy consumption and related implications can be included in Service Level Agreements (SLAs) between service users and service providers. Possible and measurable limitations could be compensated with discounts or other benefits in order to achieve high optimization and thus lower energy consumption. On the other hand users could require from service providers to use substantial level of green energy (or steady increase of it). Also relevant Operational Level Agreements (OLAs) including energy consumption aspects could be established between data centres for workload migration and access to distributed energy sources.

Due to further investigations of the scenarios and asumed use-cases the SmartenIT partners found out that overlappings of the scenarios exists. Therefore, they agreed on combination of scenarios in order to decrease the nuber of scenarios down to two. Scenario 4 dealing with energy efficiency appears horizontal, which means it will be relevant for all scenarios. Scenario 3 dealing with social awareness  will be combined with scenario 2 - global service mobility. Summarizing it up the starting four scenarios are combined to two scenarios finally:  (1) Inter-cloud communication and (2) global service mobility.

Further investigations will show how applications, like PiCsMu (presented during FIA 2013) will fit into SmartenIT.
The mechanisms RBTracker by University of Zurich, HORST by University of Würzburg, and vINCENT by University of Darmstadt will constitute ingredient parts of the SmartenIT solution.