You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@fineract.apache.org by James Dailey <ja...@gmail.com> on 2019/11/13 19:19:41 UTC
Re: Fineract with Trafodion
(CORRECTED VERSION)
Hi Rohit -
Thanks for reaching out to talk about developing a demo of Trafodion (by
Esgyn) on fineract. I think it is a worthy goal and as an open source
community we are open to any direction that has technical merit. It could
be useful for demonstrating newSQL capabilities and for fineract community
to learn about gaps we may have. Tests are good.
I think your questions are common ones - where to start and how to
understand this project. Please do read the FAQ because it has a lot of
the important context you need. If it doesn’t answer the obvious questions
then please let us know as that will help the project be better at
explaining. (https://cwiki.apache.org/confluence/display/FINERACT/FAQ)
Here are the questions I think you asked:
1. Should I based my demo on Fineract1.x or Fineract-CN ? And Where are
the relevant repos?
2. What’s the easiest path to getting a build?
3. Are there performance and regression tests to use? i.e. baseline and for
demo
And implied (I’m guessing here please clarify)
4. How well is the n-tier architecture implemented? i.e. is there a clear
data persistence layer?
5. When and where is postgreSQL vs MariaDB, etc used in the architecture?
6. How is EclipseLink used in fineract?
I’m going to try to answer these one by one, emphasis on "try". I invite
others to correct me and amplify when I get it right.
1) *FIneract1.x versus FineractCN and where are the repos*
Fineract1.x (1.4 soon to be released) is based on a code base that goes
back at least 8 years. Lots of changes but assume that the separation of
layers (concerns) is sometimes imperfect. Hibernate was removed in major
rework prior to porting to apache a few years ago. I suspect that if your
work requires any “porting” of database scheme, this will be a lot of work.
There is one repo —> https://github.com/apache/fineract/tree/1.3.0 (The
dev branch is potentially unstable.) Niklas.Urlberg gave instructions
related to deployment of this. I suspect this is NOT where you want to
spend your time.
FineractCN (cloud native version) is going through development to get to a
formal Apache release. There are some companies using a derivative of this
in production so it is real and the goal is highly scalable cloud based
deployments.
There is a repo for each Microservice, and a basic demo version requires a
subset of the available microservices.
see the wiki pages linked below
2. Instructions for build are specific for fineract1.x and FineractCN as
these are really very separate projects. Think of fineract1.x as
generation2 and fineractCN as generation3 (generation1 was a long time ago)
See —>
https://cwiki.apache.org/confluence/display/FINERACT/How+To+Build+Apache+Fineract+CN
https://cwiki.apache.org/confluence/display/FINERACT/How+To+Build+Apache+Fineract+CN
https://cwiki.apache.org/confluence/display/FINERACT/DockerHub+images
https://cwiki.apache.org/confluence/display/FINERACT/Testing
3. There are some functional and component tests for fineract1.x . These
are basic “don’t break the build” tests. There is a new working group
looking at putting together a set of performance tests - and it will base
this on existing tests done by various outside groups. I believe that Ed
Cable is convening that group soon.
4. Don’t know. Not to confuse the issue but we’ve discussed if Fineract1.x
can be easily ported to a different database. For that we say “maybe” and
“with some effort”. For fineractCN we’ve discussed if it makes sense to
have a path for other database solutions to be used - and said “not a
priority” but useful to have others suggest the way.
5.& 6.
As noted in the October 2019, the FineractCN Project got a boost from
several contributors:
https://cwiki.apache.org/confluence/display/FINERACT/2019-10+October+Report
* RDBMS migrated from MySQL to PostgreSQL by Isaac Kamga
* ORM migrated from Hibernate to EclipseLink (GSoC 2019 project by
Ebenezer Graham)
Please feel free to ask more questions. I'll try to help you zero in on the
logical next steps.
Thanks,
James
@Jdailey67
On Tue, Nov 12, 2019 at 9:53 AM Rohit Jain <ro...@esgyn.com> wrote:
> Hi Niklas and James,
>
>
>
> You have communicated with Eric Owhadi. I am the CTO for Esgyn. I
> recently integrated Trafodion with EclipseLink internally. I would like to
> see how I can create a working implementation of Fineract with Trafodion
> for demonstration purposes. Run regressions and such but also demonstrate
> high scale transactional performance. Since I am not as familiar with the
> project – with Fineract, Fineract-CN, PostgreSQL and MariaDB versions, etc.
> could you quickly guide me to where I would start – which repository etc.?
> The merits of using Trafodion could be debated. But right now I am just
> looking at seeing how well we can integrate and demonstrate differentiating
> capabilities.
>
>
>
> *Rohit Jain*
>
> Chief Technology Officer
>
> Esgyn Corporation <https://esgyn.com/>
>
>
>
RE: Fineract with Trafodion
Posted by Rohit Jain <ro...@esgyn.com>.
James,
Thanks for a very detailed and thorough response! I have enough to start exploring how to integrate Trafodion and perhaps help with the ability of FineractCN being database agnostic, although we have a vested interest in Trafodion being the scale-out distributed database choice 😊. It's heritage is Tandem Computers that was the pulse of many banking systems worldwide. I realize that as open source proponents we walk that fine line of balance between community and commercial interests.
Rohit
From: James Dailey <ja...@gmail.com>
Sent: Wednesday, November 13, 2019 1:20 PM
To: Rohit Jain <ro...@esgyn.com>; Dev <de...@fineract.apache.org>
Cc: Niklas Uhrberg <Ni...@resurs.se>
Subject: Re: Fineract with Trafodion
External
(CORRECTED VERSION)
Hi Rohit -
Thanks for reaching out to talk about developing a demo of Trafodion (by Esgyn) on fineract. I think it is a worthy goal and as an open source community we are open to any direction that has technical merit. It could be useful for demonstrating newSQL capabilities and for fineract community to learn about gaps we may have. Tests are good.
I think your questions are common ones - where to start and how to understand this project. Please do read the FAQ because it has a lot of the important context you need. If it doesn’t answer the obvious questions then please let us know as that will help the project be better at explaining. (https://cwiki.apache.org/confluence/display/FINERACT/FAQ)
Here are the questions I think you asked:
1. Should I based my demo on Fineract1.x or Fineract-CN ? And Where are the relevant repos?
2. What’s the easiest path to getting a build?
3. Are there performance and regression tests to use? i.e. baseline and for demo
And implied (I’m guessing here please clarify)
4. How well is the n-tier architecture implemented? i.e. is there a clear data persistence layer?
5. When and where is postgreSQL vs MariaDB, etc used in the architecture?
6. How is EclipseLink used in fineract?
I’m going to try to answer these one by one, emphasis on "try". I invite others to correct me and amplify when I get it right.
1) *FIneract1.x versus FineractCN and where are the repos*
Fineract1.x (1.4 soon to be released) is based on a code base that goes back at least 8 years. Lots of changes but assume that the separation of layers (concerns) is sometimes imperfect. Hibernate was removed in major rework prior to porting to apache a few years ago. I suspect that if your work requires any “porting” of database scheme, this will be a lot of work.
There is one repo —> https://github.com/apache/fineract/tree/1.3.0 (The dev branch is potentially unstable.) Niklas.Urlberg gave instructions related to deployment of this. I suspect this is NOT where you want to spend your time.
FineractCN (cloud native version) is going through development to get to a formal Apache release. There are some companies using a derivative of this in production so it is real and the goal is highly scalable cloud based deployments.
There is a repo for each Microservice, and a basic demo version requires a subset of the available microservices.
see the wiki pages linked below
2. Instructions for build are specific for fineract1.x and FineractCN as these are really very separate projects. Think of fineract1.x as generation2 and fineractCN as generation3 (generation1 was a long time ago)
See —> https://cwiki.apache.org/confluence/display/FINERACT/How+To+Build+Apache+Fineract+CN
https://cwiki.apache.org/confluence/display/FINERACT/How+To+Build+Apache+Fineract+CN https://cwiki.apache.org/confluence/display/FINERACT/DockerHub+images
https://cwiki.apache.org/confluence/display/FINERACT/Testing
3. There are some functional and component tests for fineract1.x . These are basic “don’t break the build” tests. There is a new working group looking at putting together a set of performance tests - and it will base this on existing tests done by various outside groups. I believe that Ed Cable is convening that group soon.
4. Don’t know. Not to confuse the issue but we’ve discussed if Fineract1.x can be easily ported to a different database. For that we say “maybe” and “with some effort”. For fineractCN we’ve discussed if it makes sense to have a path for other database solutions to be used - and said “not a priority” but useful to have others suggest the way.
5.& 6.
As noted in the October 2019, the FineractCN Project got a boost from several contributors:
https://cwiki.apache.org/confluence/display/FINERACT/2019-10+October+Report
* RDBMS migrated from MySQL to PostgreSQL by Isaac Kamga
* ORM migrated from Hibernate to EclipseLink (GSoC 2019 project by Ebenezer Graham)
Please feel free to ask more questions. I'll try to help you zero in on the logical next steps.
Thanks,
James
@Jdailey67
On Tue, Nov 12, 2019 at 9:53 AM Rohit Jain <ro...@esgyn.com>> wrote:
Hi Niklas and James,
You have communicated with Eric Owhadi. I am the CTO for Esgyn. I recently integrated Trafodion with EclipseLink internally. I would like to see how I can create a working implementation of Fineract with Trafodion for demonstration purposes. Run regressions and such but also demonstrate high scale transactional performance. Since I am not as familiar with the project – with Fineract, Fineract-CN, PostgreSQL and MariaDB versions, etc. could you quickly guide me to where I would start – which repository etc.? The merits of using Trafodion could be debated. But right now I am just looking at seeing how well we can integrate and demonstrate differentiating capabilities.
Rohit Jain
Chief Technology Officer
Esgyn Corporation<https://esgyn.com/>