Non-Functional Requirements ODS Big Data Working Group Proposal for criteria to which concepts shall be evaluated
2016 / 12 / 23
Concept evaluation criteria Collection of non-functional requirements
Required tools and technologies Storage levels Extensibility Privacy Security Performance Scalability Utilization of MDF formats Interoperability with external systems
Support of ODS server operations Compatibility with ODS 6 interface format Migration support Maintainability and future proof …
2
Concept evaluation criteria Collection of non-functional requirements
3
Apply to storage and access structure work packages for evaluation of design alternatives Check requirements relevance to further ODS standardization Elaborate impact based on the prescribed use cases Perform trade-off analysis (e.g. performance vs. security demands)
Concept evaluation criteria Required tools and technologies
4
Minimal demands on required tools and technologies Maximum freedom to deploy and facilitate tools and technologies List alternate compatible solution components If specific tools or technologies are required, provide proper arguments
Concept evaluation criteria Storage levels
5
Cost effective storage strategies Cold, warm, and hot storage levels Data lake(s) vs optimization for search and data processing How well is the solution adapted to the defined use cases How can the solution be generalized / standardized
Concept evaluation criteria Extensibility
6
Bring customized algorithms to the data How can engineers work with the data (without being data scientists) Flexible persistency model Integration of new data or storage types Minimal programming language restrictions
Concept evaluation criteria Security
7
Data encryption Access control
Concept evaluation criteria Performance
8
Benchmark store, update, search, retrieve, and delete data with ∙ Context data ∙ Measurement data How do data structures and formats, compressions, and redundancy impact the performance How does performance evolve with time (data growth, re-use, …) Cover huge number of smaller data chunks (e.g. from field data) and lesser number of huge data files (pre SOP validation) How does storage location and connectivity impact performance
Concept evaluation criteria Scalability
9
Propose and analyze impact of infrastructure Performance shall scale with storage and computing nodes
Concept evaluation criteria Utilization of MDF formats
10
Check options to utilize MDF files directly (w/o conversion) Map MDF data model to new storage structures Check split of MDF files for conversion and other operations
Concept evaluation criteria Interoperability with external systems
11
External data sources (e.g. weather data base or time tracking) External process tools (e.g. for visualization or reporting)
Concept evaluation criteria Support of ODS server operations
12
Map existing capabilities (session handling, search, data exchange, …)
Concept evaluation criteria Compatibility with ODS 6 interface format
13
Check options to utilize Protobuf format directly (w/o conversion) for context data or measurement data Check options to utilize CORBA / ATFX directly
Concept evaluation criteria Migration support
14
Check support for ODS 6 HTTP interface, OpenMDM, existing ODS servers (engaging CORBA with ATFX format)
Concept evaluation criteria Maintainability and future proof
15
Stability of storage model (format changes are quite costly with expected sizes) Stability of selected tools and technologies (how established are those, how well maintained) Are there open source components for the design (how active is the community) Is licensing fit for ODS BD purpose