Aggregations for charting on 500 billion documents

[goutham_pn]

Hello,

I'm working on a use case where I need to perform aggregations for charting on a 500 billion documents.

The data is related to some sales.

I have evaluated the data size as 8 mb for 24000 documents on my local machine.

Please help me in working with such data. My queries are

1. does elastic search scale for such type aggregation quries for this data size?
2. I'm looking for a response time of these aggregation quries at max as 3 sec.
3. How much of hard ware do I need to invest, any link to calculate this?

Thanks

[warkolm (Mark Walkom)]

1. What is huge?
2. Ok.
3. You need to test based on your queries and document structures.

[goutham_pn]

Hi Warkolm,

My data format is,

{
"id": "abc23",
"suId": 7,
"sDate": 1489807182683,
"cDate": 1489807182683,
"unitsSold": 9,
"revenue": 18,
"pou": 811,
"margin": 2,
"not": 6,
"woty": 11,
"zc": 599906,
"strId": 5
}

My sql quries will be some think like this
select suid, sum(revenue) where zc="599906" group by suid, sdate
select suid, avg(margin) where zc="599906" group by suid , sdate

Similar kind of queries I want to use with ES.

Please let me know if you want more information on this.