Store everything in DB instead of in-memory.

this task includes:

1. Learning how REST API works in python
2. Explore and understand the flask library in python

Explore networkx python package. Try to create a graph from a sample SONATA NSD.

• find the difference or missing values in translated descriptors (if any).

• investigate why translated descriptors (nsd and vnfd) doesnot instantiate a service in osm and sonata

• Make Postman API Collection for OSM REST calls
• Add OSM REST wrappers using the interface

this task includes:

1. List out the various components of translator engine and describe them in detail
2. create a flow diagram/chart to understand the working/flow of various components

• Make Postman API Collection for Sonata REST calls
• Add Sonata REST wrappers using the interface

Resolve Instantiation error.

Make changes to the SONATA splitting logic as suggested during the presentation.

• Establish the set up to collect metrics (NFVO capacity, VNFM capacity, latency of life cycle management of NS).

• Simulate the peak conditions and collect data.

As discussed, these are the tasks

• Collect the papers relevant to the individually assigned sections
• Add to the bibliography and cite into the sections that are relevant
• Establish the context for our implementation plan

Sent out an email to [email protected]. Waiting for the reply.

• difference between RSP and Classifier of OSM descriptor.

• connection between forwarding paths in sonata descriptor and RSP / classifier in osm descriptor.

#23 #22

Some notes on current plugin implementation - Link

Line# - Comment

37,165,170 - installing and using numpy only for random selection? not the best idea
154 - random_combination function, im sure, can be much simpler.
159 - using two lists for id and names and matching based on index? not the best idea, use dict
173 - interesting way to find the other set. but does this solution scale for n-split?
177 - can instead use dict, consider when extending support for n-split
205, 230 - hardcode. find solution for service discovery
257 - generatePackage returns false if it doesnt work. use if to confirm this (error handling)
261-263, 324 - move hardcode to config.py
271 - change HOST_URL to host
296 - to get the VIM list you can use get_vims_list() function from wrappers.SONATAClient.admin.py
335 - this needs to be supported by adaptor
318 - while doing this, can maintain a dict of all items with fields like ip, name, etc

General Comments

• Code can be a little more modular, make functions for repeated tasks (ex rest calls)
• Error handling? currently none
• Fix hardcodes, at the minimum make a config.py with all the hardcoded values (dirty fix!)
• Is it better to have a python pip package instread of having docker services? atleast in the current context
• Assumed 2-way split everywhere, nightmare for the next iteration when adding support for more.
• Add some Log.INFO for more verbosity

After the initial research of scalability, the architecture for scaling is to be made and we should come up with initial architecture choices for discussion. We will aggregate important topics of discussion and practical guides here.

Task List

• Select the approaches studied as part of initial research for further investigation (@dramesh94 )
• Come up with concrete architecture design for further investigation (@bhargavimohan )
• Do a feasibility study and investigate implementational challenges (@iashwinprasada )

Initial setup includes:

1. develop an initial logic of reading any descriptor file/data and store in memory in a common format.
2. push to github

This task includes:

1. list out all mandatory and optional parameters of OSM and Sonata Descriptors ( NSD and VNFD)
2. Compare and check if mandatory parameters in OSM is also mandatory in SONATA and vice-versa.
3. Compare and check how many optional parameters doesnot match between SONATA and OSM

• Make common interfaces of all remaining ETSI suggestions and fill excel sheet (@iashwinprasada )
• Identify all implemented calls for OSM and update excel sheet (@bhargavimohan )
• Identify all implemented calls for Sonata and uodate excel sheet (@dramesh94 )
• Review, finalize and distribute implementation (@iashwinprasada )

Study about JSON/ YAML file processing and different APIs which can be used to extract all the information from the files.

Pulling out NS elements from the NSD and make a map out of it which is to be forwarded to graph processor.

@bhargavimohan Tasks:

• VM Config and Ubuntu setup

• Install tools

  • OSM
  • netdata
  • openstack

• Decide NSD/VNFD

  • Have multiple options (2-3)
  • Establish baseline experiment (dummy nsd/vnfd?)

• Decide experiment data structure

  • what to collect?
  • how much/long to collect? (time?)
  • charts?
  • Questions to be answered by the test data?
  • How to store and how to present results?

• Start experiment

  • collect data as planned
  • document results

• Repeat experiment with different NSD/VNFD

  • Also vary other parameters (# instances?)
  • Try inducing some load on the running VNF

• ....

@dramesh94 Tasks:

• VM Config and Ubuntu setup

• Install tools

  • Pishahang
  • netdata
  • openstack

• Decide NSD/VNFD

  • Have multiple options (2-3)
  • Establish baseline experiment (dummy nsd/vnfd?)

• Decide experiment data structure

  • what to collect?
  • how much/long to collect? (time?)
  • charts?
  • Questions to be answered by the test data?
  • How to store and how to present results?

• Start experiment

  • collect data as planned
  • document results

• Repeat experiment with different NSD/VNFD

  • Also vary other parameters (# instances?)
  • Try inducing some load on the running VNF

• ....

Study NSDs (VNFFGDs) of following frameworks:
OSM
Sonata
OpenBaton

We will be defining functional blocks of the WP2.