CCSDK/SLI usage and further updates questions #ccsdk #sdnc

Ruslan Kashapov

Hi team

My name is Ruslan. I'm a software engineer at, also participating in ONAP CPS project.

Currently i'm checking the state/usage of CCSDK/SLI framework and I'd like to clarify a few questions (below)
related to the subject. 
Will appreciate any reply, brief answers are also ok.

1. The very first mention of SLI i found is SDNC project proposal document dated 2017; all the SLI core related activity
(according to repository change history) is also mainly references 2017-2018. 

- Is the framework abandoned and no longer supported or there some other reasons there are no recent updates?

- Are there any plans (or demand) for update/refactor existing code in order to prolong the framework lifetime
 (implement new features, fix existing bugs, reliability and/or performance improvement etc)?

2. The direct graph as a flow descriptor is quite complicated (unusual, non-standard) and requires significant
time investment in order to understand start using it. So regarding the approach in general

- Is the usage of direct graph reliable/effective enough (historical perspective, sort of a usage feedback)
  to justify the complexity of actual flow setup?

- Is the existing model (xsd), set of available nodes, design in general is sufficient to describe the desired flows? 
  (imo it seems impossible to reach same node using different paths)

- Is it makes sense to improve (evolve) the direct graph as flow descriptor? Or it's easier to bring (implement)
 completely new approach (e.g. like in openstack mistral workflow) then fix existing issues?

- Are there any plans for direct graph flow migration to some alternative?

3. The only tool i found to work with direct graphs is dgbuilder. However it lucks some core features (like 

loading existing graph from xml file for editing) and it's interface is quite messy.

- Is dgbuilder being updated/supported? 

- Are there any plans regarding the tools (fixing current, move to alternative, etc)?

4. I noticed the activities on decoupling the SLI from SDNC (ODL based) to a separate SpringBoot based
application -- the sliboot (ccsdk/apps) and gra-app (sdnc/apps). Seems GR API relies on SLI framework,
and it a bit contradicts the first impression the SLI is being in abandoned state.

- Is the SLI being decoupled in order rid of dependency on ODL? or in order to isolate the existing SLI
 based functionality with subsequent replacement to something more modern/ reliable/ other?

- Is it expected the decoupled application to be used for southbound communications, or it's mainly for 
  integration with other ONAP components like A&AI etc?

- Are there any plans to move SLI based (Direct Graph execution) application into a separate generic use project?



Ruslan Kashapov

Join to automatically receive all group messages.