On Dec 11, 2010, at 5:34 PM, Rudi Ahlers wrote:
> On Sat, Dec 11, 2010 at 6:27 PM, Joe Landman
> <[email protected]> wrote:
>> On 12/11/2010 11:17 AM, Rudi Ahlers wrote:
>> [..]
>>
>> --
>
>
> True, I fully agree with you on that point. I myself don't like vendor
> lock-ins. But, I do like the simplicity that off-the-shelf NAS devices
> offer, as apposed to a DIY one. And they often offer more tools than a
> normal Linux box with iSCSI as well.
>
hi all,
[Disclaimer: I work for Scality]
Rudi, first and foremost, i think you have to pinpoint the main characteristic
you want to improve on your system. It sounds like it is scalablity. Beyond
NAS, there is an array (no pun intended) of choices for you to consider
(scale-out NAS, SANs, clustered file system, dispersed storage ...).
But to figure out the best system for you, one needs to take a hard look at the
kind of data you're storing and at the access patterns and requirements for
those.
For large systems (meaning > 100 TB and/or lots of files (of whichever size)),
getting rid of the filesystem layer is sometimes very efficient. Hence, one
should think about Object Storage as a solution. They bring high-reliability
and durability (through self-replication, self-healing) with cost-efficiency
using commodity hardware. It is the technology used by lots of public cloud
vendors to offer their service for cheap and still be profitable. Those guys
inherently need to be able to scale almost infinitely (look at Amazon or
Rackspace).
If you are dealing with unstructured data as opposed to a relational database
for example; if there are millions and millions of objects that you need to
access quickly, object storage might be for you.
Rudi, what do you actually store ? for what kind of service is the storage
layer used ? Are you storing emails/backups or hosting your employees' file
sharing service ? Basically, if you're storing any kind of data which size you
know is likely to grow massively and you want to be able to:
1 - allow it
2 - afford it
3 - operate it at the lowest possible cost
, then i would strongly suggest you look into object storage technology (there
are a couple of opensource options as well as vendor solutions).
At scality, we have developed such an object store which scales smoothly up to
petabytes with off-the shelf servers logically brought together in a ring.
While other solutions' performance usually degrade with time, our performance
is similar to a high-end SAN from the start and stays roughly the same as we
scale up to petabytes.
We gracefully scale in capacity and performance by just adding nodes to the
system (without service interruption), so we're never limited by a box design
(either in maximum nb of drives or by network capacity).
If you feel like you could use object storage to store your data, please have a
look at our technology and get in touch - http://bit.ly/fY6eMm
Happy Holidays !
-Marc Villemade
http://linkd.in/heve30
_______________________________________________
Gluster-users mailing list
[email protected]
http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
