Hi my name is Nigel Williams and I’m a
Solutions Architect for Hewlett Packard Enterprise and today we are going to
talk about scaling within Nimble Storage now one of the most regular questions I
get when customers look at Nimble as an option, is how they can scale us up and
move on to their expanding requirements. Thankfully it’s an extremely simple
process to do now judging by this table here what I’ve done is started with a
very base model 20 now that can be an AF 24 all-flash or it could be a hf 24 the
hybrid luckily the rules are the same so if we look at this environment we’ve got
this base model to start off with just a head no shelves no it grades and we find
that what we need to do is there’s been an increase in performance requirement
what we can do is quite happily move up to the 40 controller we can move up
again to the 60 and if it’s an all flash we can even go up again to the AF 80 now
the reason we would do this is that we have a much li greatly increased
performance requirement and through CPU but the beauty is that we don’t have to
go in stages if you have a very large increase you can go straight from a 22
and 80 and the or flash or from a 20 to the 60 in the hybrid fashion with no
downtime and that’s a major benefit when to look at when choosing a solution if
you have this large performance requirement you can simply order the
upgrade controllers whether as part of our timeless storage program or just
through a regular purchase swap them out on disruptively
and gets that extra performance there and then when you need it now a far more
common option is to grow capacity whether it be through growth over time
growing needs the business or just the ever generation of data that we have now
luckily again with nimble we’ve got multiple options if you are in the
hybrid environment we can go all the way up to six shelves now these do not need
to be identical we can go in stages of 21t all the way up to two hundred and
ten terabytes with 15 terabyte drives on the way this means that you have a
massive amount of raw addressable capacity available per chip per array to
the point where we can literally go six times 210 terabyte shelves
raw now on an all flash rules are slightly different the disks are denser
and we have more disks for shelf so that gives us an ability to add an extra two
shelves but in regarding of the amount of usable capacity and data reduction we
are still there with what we need so we can easily scale performance and we can
easily scale capacity both non disruptively
and both coming online within less than half an hour again no downtime required
no maintenance window required this enables you to be extremely flexible in
your IT operations and move forward with other projects instead now you might
notice I’ve forgotten a single Arab so what the scaling here is what is scaling
out now we’ve scaled at the performance as much as we can do in a single in a
single array so of course what we want to do is to go to multiple arrays
luckily nimble has the ability to cluster not only two but four heads
together now the beauty of this is that they do not need to be the same I can
quite happily blend and all flash and a hybrid together or I can have four of
the same in if I just purely want more performance giving me an exceptional
amount of CPU capacity to use again the beauty of this is that it doesn’t tie
down any of the rules if I want six shelves on a hybrid I can then multiply
that by four to give me 20 for expansion shelves under a single management GUI
allowing me to manage over a petabytes of storage now if that isn’t enough for
most workloads I’m not sure what it so in this environment you have the ability
to scale performance scale capacity and scale outward to a cluster long
disruptively if you want to find out more please feel
free to go to our website hbo.com for storage

HPE #Nimble Storage Scale-to-fit Architecture

Leave a Reply

Your email address will not be published. Required fields are marked *