# \_uid size\_in\_bytes optimization
**URL:** https://discuss.elastic.co/t/-uid-size-in-bytes-optimization/90029
**Category:** Elasticsearch
**Created:** [June 20, 2017, 4:44am UTC](https://discuss.elastic.co/t/-uid-size-in-bytes-optimization/90029 "2017-06-20T04:44:10Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@bCast](https://discuss.elastic.co/u/bCast)
#### Post date: [June 20, 2017, 4:44am UTC](https://discuss.elastic.co/t/-uid-size-in-bytes-optimization/90029/1 "2017-06-20T04:44:10Z")
Hi! I use elasticsearch 5.2.1 to store, search and aggregate timeseries data of network connections. Index is about 35Gb per day, single type and contans no more than 10 fields which are strong typed. When I analyzed segments info of an index I found that most memory usage (size\_in\_bytes) belongs to \_uid field which I do not use directly.
I tried to play with id generators but do not see much savings. So is there any plan to optimize \_uid field heap memory consuption (for example move it to numeric form) or completely disable it?
---
### Author:  [@polyfractal](https://discuss.elastic.co/u/polyfractal)
#### Post date: [June 20, 2017, 8:54pm UTC](https://discuss.elastic.co/t/-uid-size-in-bytes-optimization/90029/2 "2017-06-20T20:54:29Z")
Yep, it's actively under discussion (a blocker for 6.0 actually). See: [https://github.com/elastic/elasticsearch/issues/25240](https://github.com/elastic/elasticsearch/issues/25240) and [https://github.com/elastic/elasticsearch/issues/11887](https://github.com/elastic/elasticsearch/issues/11887)
---
### Author:  [@system](https://discuss.elastic.co/u/system)
#### Post date: [July 18, 2017, 8:54pm UTC](https://discuss.elastic.co/t/-uid-size-in-bytes-optimization/90029/3 "2017-07-18T20:54:30Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.