BigData (Hadoop/NoSQL)

[Muhammad Nuzaihan Bin Kamal Luddin]

And sorry for keeping it to myself for almost a year.

There you have it. swapcache is an interesting take as it is different 
than L2arc.

It's either to port Hadoop/HBase to C or run JVM in kernel-space. The 
advantages of running in JVM is not really technical, it's just that 
they would have more developers and most BigData tools are in Java.

Regards,
Muhammad Nuzaihan

On 30/10/2013 10:05, Muhammad Nuzaihan Bin Kamal Luddin wrote:
> Hi All,
>
> I had this idea since the early 2013 - even before "osv.io" came out.
>
> Looking at the characteristic of DragonFly, it looks more modern for 
> really special applications and HAMMER looks an interesting step for 
> Desktops as well.
>
> - purpose was for BigData optimisations (Hadoop) and NoSQL
> - the overhaul of a typical UNIX design was more interesting with 
> DragonFlyBSD
> - DragonFlyBSD project just has acquired a 48-core opteron machine to 
> stress test their performance gains
> - swapcache (similar to bcache on Linux and L2arc on ZFS)
> - HAMMER filesystem (and the under-development HAMMER2) with an 
> interesting snapshot feature. (to solve BigData storage (archival) 
> problems)
>
> Note: a typical Data scientist does not require such a titan - they 
> only need a simple relational DB (MySQL/MSSQL) and a normal Windows 
> Laptop.
>
> Regards,
> Muhammad Nuzaihan
>

-- 
Taqi Systems
269-J Jalan Panji
Kampung Chempaka, Kota Bharu, Kelantan 16100

Phone: +6097738657
Mobile: +60199788657

pub   4096R/4C77F88C 2013-04-06 [expires: 2015-04-06]
       Key fingerprint = 2FE1 87FA E775 2E05 CC0F  B3F6 3CB7 C65F 4C77
F88C
uid                  Muhammad Nuzaihan Bin Kamal Luddin
<muhammad at taqisystems.com>