Janus Graph as replace for NOSQL database in Web Applications for non-bulk randomly generated data


Jane <shabs...@...>
 

So by separating the data into two different databases, traversals on vertex properties would not be possible without external look ups. 

Would there be any specific optimizations that I could possibly do since I would like to also be able to use OLAP part of Janus. 


On Tuesday, June 20, 2017 at 3:29:16 PM UTC-4, Robert Dale wrote:
You could keep your schema-less data in your existing NoSQL db and then add a graph database for storing relations.

Or try a multi-model graph database. There are several listed here http://tinkerpop.apache.org/#graph-systems

There's no one right way to do it.

Robert Dale

On Tue, Jun 20, 2017 at 2:32 PM, Jane <sh...@...> wrote:
Can I/Has anyone use Janus Graph as a replacement for NOSQL databases in web applications? I'm talking about high throughput on both reads and writes (non-bulk, randomly generated by user traffic). A lot of examples that are getting tossed around here are just a bunch of people exporting their data from somewhere else and bulk importing it into Janus Graph for analysis. 

It seems that a blocking issue for me with Janus Graph is that for it to remotely even perform decently it requires an index and schema to be pre-generated for the data set. With my use case it is simply not possible to determine everything that a user might want to possibly upload to the database and hence why we use a NOSQL database. However, the data is also highly relational and would benefit from being stored in a graph database. 

Has anyone gotten over this hurdle or have any advice? 

--
You received this message because you are subscribed to the Google Groups "JanusGraph users list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to janusgraph-use...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Robert Dale <rob...@...>
 

You could keep your schema-less data in your existing NoSQL db and then add a graph database for storing relations.

Or try a multi-model graph database. There are several listed here http://tinkerpop.apache.org/#graph-systems

There's no one right way to do it.

Robert Dale

On Tue, Jun 20, 2017 at 2:32 PM, Jane <shabs...@...> wrote:
Can I/Has anyone use Janus Graph as a replacement for NOSQL databases in web applications? I'm talking about high throughput on both reads and writes (non-bulk, randomly generated by user traffic). A lot of examples that are getting tossed around here are just a bunch of people exporting their data from somewhere else and bulk importing it into Janus Graph for analysis. 

It seems that a blocking issue for me with Janus Graph is that for it to remotely even perform decently it requires an index and schema to be pre-generated for the data set. With my use case it is simply not possible to determine everything that a user might want to possibly upload to the database and hence why we use a NOSQL database. However, the data is also highly relational and would benefit from being stored in a graph database. 

Has anyone gotten over this hurdle or have any advice? 

--
You received this message because you are subscribed to the Google Groups "JanusGraph users list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to janusgraph-users+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Jane <shabs...@...>
 

Can I/Has anyone use Janus Graph as a replacement for NOSQL databases in web applications? I'm talking about high throughput on both reads and writes (non-bulk, randomly generated by user traffic). A lot of examples that are getting tossed around here are just a bunch of people exporting their data from somewhere else and bulk importing it into Janus Graph for analysis. 

It seems that a blocking issue for me with Janus Graph is that for it to remotely even perform decently it requires an index and schema to be pre-generated for the data set. With my use case it is simply not possible to determine everything that a user might want to possibly upload to the database and hence why we use a NOSQL database. However, the data is also highly relational and would benefit from being stored in a graph database. 

Has anyone gotten over this hurdle or have any advice?