6

Hi I looking for In memory data grid or similar one.

My use case.

  1. Data griding in memory. scale out available.
  2. backup node available.
  3. persistent backup available.
  4. (optional) free or opensource solution

I did googling and I found candidates below
- Apache Ignite
- Redis cluster
- Hazelcast(community)

I prefer Ignite to Hazelcast because, Ignite support use direct buffer.

But I don't know Redis cluster partitioning whether it is stable or not. and, I don't know if apache ignite performance better than redis cluster or not.

Apache Ignite comparable to redis cluster? or impropert comparison? Thanks for your answer

moon
  • 165
  • 1
  • 3
  • 6

3 Answers3

7

But I don't know Redis cluster partitioning whether it is stable or not

Redis cluster feature is stable since 3.x version and used in production by many companies.

Apache Ignite comparable to redis cluster? or impropert comparison?

Comparison Apache Ignite vs Redis only is wrong, because these projects have different grade. Redis is positioned as a storage and not as a Data Grid like Apache Ignite. So for proper comparison Apache Ignite should be compared vs Redisson - Redis Java Client with features of In-Memory Data Grid. It offers the same features as Apache Ignite.

Redisson supports fully managed Redis services like AWS Elasticache, Azure Redis Cache. You don't need to manage/deploy/maintain Redis cluster by yourself of hire devops to do this. Apache Ignite doesn't offer such feature and you should manage/deploy/maintain it by yourself.

Nikita Koksharov
  • 10,283
  • 1
  • 62
  • 71
2

I used Redis in production for one of the largest US mobile network operator (IoT department). It is stable from 2.8 (Master/Slave) but cluster stable is from 3.2. Used 2.8 for 3 years and 3.2 cluster for 2 years on production with about 50k TPS load with no restart for years and no issues (except BGSAVE and memory issues but that was due to RAM limitations).

If we compare Redis and Apache Ignite:

  1. Performance. Redis is faster, single thread and 100% in memory.
  2. Data structure. Redis is key-value storage (even that is not a limitation, you can imagine and map almost everything in key-value models). Ingrid is a data grid as it was mentioned above.

If you are looking for a memory data grid and performance is on second priority then Ingrid will be more appropriate for you.

Denys
  • 1,308
  • 12
  • 13
  • Can below thing achievable in Redis Cache? Let's say I have 100 records. Cache can hold only 40 records (LRU) and 100 records on disk file(Not in any other database) So if anything from these 100 records are requested I never have to go actual database(for e.g. Sybase db)? If key found in 100 records but its not present in in-memory cache(40records) then get that key, put in in-memory cache and swap out other key using eviction policy to a disk file(on disk i have 100 records always) If key not present in Cache and disk file then only we have to go database (Sybase) – Gopu_Tunas Sep 10 '19 at 07:21
  • Ignite is faster then redis, in general – frank-dspeed Mar 11 '20 at 22:49
1

Redis only provides a key-value storage, while Ignite is much more functional. Here is a good feature comparison provided by GridGain: https://www.gridgain.com/resources/product-comparisons/redis-comparison

Which one to use, depends on your requirements and expectations.

Valentin Kulichenko
  • 8,365
  • 1
  • 16
  • 12
  • Thanks. I only use key value storage – moon Feb 01 '17 at 00:53
  • 1
    I think this comparison is incorrect because they should compare GridGain to other Redis based Data Grids like [Redisson](https://redisson.org) and not Redis itself. Redis is a low level storage. – Nikita Koksharov May 04 '17 at 15:10