Hi,

Ignite does introduce some overhead to every entry, mostly because of its
distributed nature. This is a tradeoff for having all the features provided
by Ignite. 200 bytes per entry sounds correct, not taking indexing into
account. See this page for details:
https://apacheignite.readme.io/docs/capacity-planning-bak

-Val


On Tue, Dec 6, 2016 at 2:45 AM, rishi007bansod <[email protected]>
wrote:

> I have loaded table consisting of 150000 entries in cache. From heap dump I
> understood that, byte[] array stores key & values in this table.
> But following objects also gets added with each entry,
>
> BinaryObjImpl = 40 Bytes *150000 * 2(key+value) = 12 MB
> GridAtomicCacheEntry = 64 Bytes * 150000 = 9.6 MB
> Also, I am getting* per entry 200 Bytes overhead = 200 * 150000 = 30 MB*
>
> Also, I observed that ConcurrentSkipListMap$Node,
> ConcurrentSkipListMap$Index, GridH2ValuecacheObject,
> GridH2KeyValueRowOnHeap
> consumes about 29 MBs
>
> How can I reduce, above mentioned number of objects(As Byte array(key,
> value
> pairs) size is only 139 MB where total consumption is 232MB)
>
> <http://apache-ignite-users.70518.x6.nabble.com/file/n9412/1.png>
>
>
>
>
> --
> View this message in context: http://apache-ignite-users.
> 70518.x6.nabble.com/Memory-Overhead-per-entry-in-Apache-Ignite-tp9412.html
> Sent from the Apache Ignite Users mailing list archive at Nabble.com.
>

Reply via email to