Use cases for Ignite

[Prachi Garg]

Hi,

I've been working on putting some use cases for Ignite, on the Ignite
website, and would like to get some suggestions from the community. What
can be other suitable use cases in addition to the following?

*Caching*

   - Database Caching
   - Hibernate L2 Cache
   - Web Session Clustering


*Data Grid Replacement* (provide comparisons with other In Memory Data Grid
applications)

*Ignite for Spark*

   - Shared Storage for Spark Tasks
   - Faster SQL


Thanks,
-Prachi

[Konstantin Boudnik-2]

Two more just from the top of my head:
  - distributed file system (IGFS) w/ optional caching for HDFS-like file systems
  - in-memory mapreduce and acceleration for Hive and similar applications

Cos

On Tue, Aug 04, 2015 at 01:05PM, Prachi Garg wrote:

> Hi,
>
> I've been working on putting some use cases for Ignite, on the Ignite
> website, and would like to get some suggestions from the community. What
> can be other suitable use cases in addition to the following?
>
> *Caching*
>
>    - Database Caching
>    - Hibernate L2 Cache
>    - Web Session Clustering
>
>
> *Data Grid Replacement* (provide comparisons with other In Memory Data Grid
> applications)
>
> *Ignite for Spark*
>
>    - Shared Storage for Spark Tasks
>    - Faster SQL
>
>
> Thanks,
> -Prachi

[dsetrakyan]

Inline...

On Tue, Aug 4, 2015 at 1:44 PM, Konstantin Boudnik <[hidden email]> wrote:

> Two more just from the top of my head:
>   - distributed file system (IGFS) w/ optional caching for HDFS-like file
> systems
>   - in-memory mapreduce and acceleration for Hive and similar applications
>

Agree. I think the titles are too long, however. How about:
- In-Memory File System as HDFS Cache
- Faster Hive with In-Memory Map Reduce

>
> Cos
>
> On Tue, Aug 04, 2015 at 01:05PM, Prachi Garg wrote:
> > Hi,
> >
> > I've been working on putting some use cases for Ignite, on the Ignite
> > website, and would like to get some suggestions from the community. What
> > can be other suitable use cases in addition to the following?
> >
> > *Caching*
> >
> >    - Database Caching
> >    - Hibernate L2 Cache
> >    - Web Session Clustering
> >
> >
> > *Data Grid Replacement* (provide comparisons with other In Memory Data
> Grid
> > applications)
> >
> > *Ignite for Spark*
> >
> >    - Shared Storage for Spark Tasks
> >    - Faster SQL
> >
> >
> > Thanks,
> > -Prachi
>

[dsetrakyan]

Thanks Prachi,

I just have a few comments below...

On Tue, Aug 4, 2015 at 1:05 PM, Prachi Garg <[hidden email]> wrote:

> Hi,
>
> I've been working on putting some use cases for Ignite, on the Ignite
> website, and would like to get some suggestions from the community. What
> can be other suitable use cases in addition to the following?
>
> *Caching*
>
>    - Database Caching
>    - Hibernate L2 Cache
>    - Web Session Clustering
>

I think we should also add:

- JCache Provider


> *Data Grid Replacement* (provide comparisons with other In Memory Data Grid
> applications)
>

Given that Ignite Data Grid is already replacing many commercial and some
open source data grid products, I think an unbiased comparison outlining
main differences does make sense. If I can suggest, we can have:

- Hazelcast
- Redis
- Coherence
- GemFire


> *Ignite for Spark*
>
>    - Shared Storage for Spark Tasks
>    - Faster SQL
>

How about:

- Shared Memory Layer for Spark
- Faster Spark SQL with In-Memory Indexes


>
>
> Thanks,
> -Prachi
>

[pooja3490]

Hi,

I wanted to know that in use case where we need in-memory big data for on-the-fly computation, does Ignite gives better performance as compared to kdb (which is also an in-memory database)?

As far as I understand, Ignite is used for caching database that store data on disk or to improve performance of systems that already use spark/hadoop.

It would be very informative if someone could share comparison of Ignite Vs. KDb keeping the above use case in mind.

Thanks

[dsetrakyan]

On Sun, Jul 31, 2016 at 11:50 PM, pooja3490 <[hidden email]> wrote:

> Hi,
>
> I wanted to know that in use case where we need in-memory big data for
> on-the-fly computation, does Ignite gives better performance as compared to
> kdb (which is also an in-memory database)?
>
> As far as I understand, Ignite is used for caching database that store data
> on disk or to improve performance of systems that already use spark/hadoop.
>
> It would be very informative if someone could share comparison of Ignite
> Vs.
> KDb keeping the above use case in mind.
>
>

We don't have a comparison of Ignite to KDB, but at a high level KDB is a
time-series database. It also does not seem like KDB is a distributed
system either.

If you are interested in time-series queries, then in Ignite it would be
achieved using SQL and querying *between* timestamps using standard SQL
syntax.

Ignite performance should be better simply because Ignite can split a data
set into multiple pieces and run the query in parallel across multiple
nodes. So, if you have 10 nodes, then each node should work in parallel
with D/10 data size, where D is your total data size. Given that each node
has to process less data, queries should execute faster.


>
>
> --
> View this message in context:
> http://apache-ignite-developers.2346864.n4.nabble.com/Use-cases-for-Ignite-tp2132p10290.html
> Sent from the Apache Ignite Developers mailing list archive at Nabble.com.
>