Reply to: Best paying database developer environment?

[AtW]

Sybase to SQL Server might be a lot easier to move.

Sybase to HADOOP will require changing a lot more, since it's not like for like. Chances are HADOOP _might_ be more appropriate/cheaper, but all depends on what you actually doing.

We do use Postgres, seems ok, but 10 TB data is pretty heavy.

[Platypus]

Interesting https://www.theregister.co.uk/2017/0...l_on_big_data/

[Brummie]

Hadoop has been discussed and is very fashionable. The main objection seems to be we only have a 10TB database so are not big enough for big data. Though 10TB seems quite large to me - clearly I need to get with the times.

It's not just the Volume that should put you off from using it. Considering that Data is the new oil :-), Do you expect to store and process Unstructured, Semi-structured data (e.g Audio/Video streams, Emails etc)? Do you need transactions (Big data doesn't support transactions)? Do you need real-time analytics? If your use case is Geo then Mongo has a fantastic support for distance calculations and it is way faster than anything I have seen or Do you need to know who is following whom (degrees of separation!) then go with Neo4j, its node traversal is way faster than anything else.

[SeanT]

You Are Not Google

"Did you know you can buy a terabyte of RAM for around $10,000?" - a month, if using the public cloud...

[Sysman]

Hadoop has been discussed and is very fashionable. The main objection seems to be we only have a 10TB database so are not big enough for big data. Though 10TB seems quite large to me - clearly I need to get with the times.

You Are Not Google

[MrMarkyMark]

I get these emails everyday....

Hi Ya Dim, I hope you are well.
My blessed company has just renewed contract, inshallah, this is good doings as we will working together for a very, very, long time.
Shiva has risen from dead. We have managed to revert database and analysis cubes to a previous running state from corporate back up. Nothing will run, do you know anything about indexing doings? Please revert at the soonest.

[DimPrawn]

Quote Originally Posted by BobShawadiwadi View Post
Hi ya Dim, after some Production doings all the tables have gone the way of Shiva,
as database expert can you please revert with the step by step necessary doings and the way to fix new non functioning back up. Thanks

I get these emails everyday....

[MrMarkyMark]

All DB work has gone to India. You're a decade too late.

Hi ya Dim, after some Production doings all the tables have gone the way of Shiva,
as database expert can you please revert with the step by step necessary doings and the way to fix new non functioning back up. Thanks

[fool]

MongoDB. For the times when you don't really care if your data is there or not.

[DimPrawn]

All DB work has gone to India. You're a decade too late.

[darmstadt]

I've been testing DB2 BLU with analytics, Postgres, MongoDB, MariaDB along with weird stuff like Redis, Spark, Cassandra, Geode, etc but it really depends on what you want to do with it. DB2 BLU was not bad and Postgres is nice and simple (R/W was slow though)

[Brummie]

I would ask myself this. Should it support 3V's - Volume, Velocity and Variety. If yes, go BigData stack. If you are looking at other DBs then decide if they should support Graphs (e.g. GraphDB) or Relational?
If relational, then Do you need stored procedures, triggers etc.

Your use case would say which one is suitable.

[eek]

Mongodb serves certain use cases. It's great but it's overkill for some purposes. sql server and azure will open a lot more doors

[BigRed]

+1

Azure and hybrid are both good options. SQL Server was based on Sybase so should be an easier transition. MongoDB is good for some things but it's very immature in other areas.

[quackhandle]

For career opportunities SQL server hybrid (cloud and on-prem)..

This. But if you want to be all trendy suggest MongoDB.

qh