geode-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gregory Chase <>
Subject Re: Geode wan vs Kafka replication
Date Thu, 21 Feb 2019 18:43:20 GMT
Hi Aashish,
I spend a lot of time thinking about this for other systems.  Here's my
thought regarding Geode...

You want both, depending on the use case.

If you are only dealing with data being replicated and shared between Geode
clusters, use WAN replication.

If you want to make messages, data, or events available for ANY other
systems, use Kafka.

Where it gets interesting is when your Geode clusters serve different use
cases.  For example, one handles primary transactions, and perhaps the
other is used for analysis, or as DR.  WAN replication is probably still
the easiest, but if other kinds of systems will want the results of
application transactions, then feeding messages into Kafka will give you a
very adaptable IT infrastructure.


On Thu, Feb 21, 2019 at 10:08 AM aashish choudhary <> wrote:

> Hi,
> Looking for some inputs over here. We are looking to use geode wan
> replication for data transfer across regions(Nam to Apac to Emea). We have
> run couple of tests(large datasets too) with Geode wan replication and in
> principle it seems to work ok. Another option is to look for distributed
> messaging platform like Kafka.
> Couple of questions
> Would geode be able to scale well compare to Kafka over the long run?
> Performance Geode wan vs kafka. Anything particular that we should look at
> before confirming the solution.
> With best regards,
> Ashish

Greg Chase

Business Development

Pivotal Software


View raw message