Objective 3: Make the HD-service lifecycle safe, predictable and consistent
Summary: This objective includes providing a tool to HD-service developers to support them mastering development of HD-services. Guidance should be provided to help designers deciding how to properly break down the business logic and where to distribute it, taking into account the capabilities and specificities of the individual nodes of the heterogeneous computing infrastructure. Validation and verification support should also be provided to certify certain properties of the system prior to deployment. Targeted properties include functional correctness but also extra functional properties critical to HD-Services like performances and data integrity and privacy.
Challenges: This objective is challenging as it requires to reason on an adequate abstraction layer, which on the one side includes necessary details about the system and platform to make acceptable assumption and useful checks, still on the other side is abstract enough to make such validation tractable and portable to different platforms and applications. Furthermore, to make statements before deployment means that either the analysis has to be made statically on the models or the physical nodes, potentially very numerous, need to be mocked.
Approach: For the early verification of properties such as response times, integrity and privacy of data, the HD-Service models will include information on Quality of Service (QoS) requirements. For example, the maximum time between an event and a system reaction to it, the level of reliability/security of a particular communication link or the level of confidentiality of a particular piece of data. This information will be exploited for the validation of each particular deployment in order to ensure that requirements are met. The approach will use generative techniques to generate the code able to mock physical devices and leverage the elasticity provided by the cloud to instantiate a large number of these mock-ups in order to simulate real life deployments. We will then use model-based testing to generate tests (unit, integration and load) and execute them on this simulated infrastructure. The use of simulation and mock-up will allow going gradually from a simulation of the HD-Service to its final deployment by replacing simulated nodes by actual nodes as they become available. The simulation will allow testing the overall logic of an HD-Service early on (before the actual infrastructure is available) and facilitate integration testing by allowing the introduction of nodes and platforms one by one into the system.
- A set of tools that complement the HEADS IDE with checkers and simulators
- A metamodel to specify test constraints and objectives, including QoS
- A set of tools to generate test benches and their infrastructure
- A representative set of sample test benches
Tellu participating at Vitalis 2017May 02
Tellu participated at Vitalis 2017 in Gothenburg, Sweden, on 25th-27th of April 2017.
TelluCloud innovation developed in the HEADS project at EC's stand in CEBIT 2017Mar 20
We are very glad to announce that Tellu AS has officially been selected as one of the top inn