So, predictive analytics has been around
for a decade already - and today everybody seems to have
their own implementation. Dell EMC has Cloud IQ, - Pure Storage has Pure1 Meta - NetApp has Active IQ etc etc. But today - we are going to take a closer look at HPE InfoSight - which still seems to be the benchmark - for predictive analytics
capabilities in 2020. Stay with me! Hello! it's Markus again. So, if you are new to the channel - Welcome! This is where we talk about everything
around enterprise IT. Cloud, servers, storage, networking, virtualization - all that fun stuff! If you haven't subscribed yet - hit that subscribe button - and don't forget to click that little
bell button as well - to get the notification of my new videos. actually just before I started filming this - I noticed that I broke - the 500 subscriber barrier! thank you every single one of you! This keeps me going and releasing
even more videos in 2020. So, thank you, thank you, thank you! Now, with that out of the way
let's get down to business! Okay, so let's start this
from the end-user point of view. so we have an end user - who wants to access some data
somewhere. So, he only wants to use that data, analyze it,
whatever he wants to do with that. But it is not that simple. You cannot directly access the data. We humans don't speak binary very well - so, we need some assistance. What we need is some kind of end user device. For example a mobile phone - with an antenna obviously... 2020, right? And the mobile phone is then accessing some
applications most of the time. So mobile phones are running,
accessing some applications - and applications are then visualizing - the data and accessing the data for us. What apps are running on top of? Most of the time still - on hypervisors or containers
nowadays as well. hypervisors most of the time are running
on top of an operating system. Operating system on the other hand is - installed on some hardware: servers. And servers as we know they
have many components. We have motherboards, we have CPUs - we have memory, - we have different kinds of PCI cards - of which one important one
would be a network adapter. ...or a HBA card. The network adapter is
connected to a switch - which has another transceiver there
and some cabling to do. That switch has many other ports - some of which are connecting to - our storage array. and this is where we finally have our disks - Where the precious data is stored in. So, what we need to do in order for this - end-user to read the data? We first need to request it. We need to go through the end user device - application, hypervisor, some drivers, - some operating system services. Then we go to server - where we find firmware - we find mother board, memory, PCI cards - buses, network adapter, transceiver, cable, transceiver - switch, another switch, - then we finally arrive to array - where we have transceivers,
we have the backends - we have more cabling and so on. And then finally our request to get the data - arrives to the disks. And guess what? Then we have to travel back the
same route with the data! All the way from here up there. And there's plenty of components on the way - which a) can break down - or b) just slow down the traffic. They were calling this problem - this specific problem that there's something
problematic in that stack - they started to call that App-Data Gap. So, a gap between applications and data. If end-user reports if there's something - wrong with their application or
accessing the data - then starts the long troubleshooting
of the problem. You have to first of all track down the problem. Comparing logs from arrays, from switches
and compute resources - operating system, everything
like that. Nimble Storage came up with the
idea of InfoSight. Some kind of automation system - that helps system administrators in
their tremendous tasks. How InfoSight then works? We have a Nimble Storage array first of all here - which is running a Nimble OS - as it's operating system. And as a part of Nimble OS - there is this one process let's say - running in the operating system that's
constantly monitoring - the disks - storage array itself - the health status, all the temperatures,
all the throughputs, - the spinning disk statuses etc etc. Monitoring the storage array itself - is nothing too different to what everybody
else is doing on the market - DellEMC, NetApp, Pure Storage, IBM, HPE... everybody is doing this. And storage arrays in general
are quite well known - to be quite good reporters of their
health status and performance. So, nothing too different here. Vast majority actually of the problems - occur somewhere else but
storage array - although they might seem
like storage problems. So, if we want to have a comprehensive
view of the problem - we actually have to understand
the whole environment - or as much of it as possible. Not only storage array. And that's why InfoSight actually visibility - to outside storage. We can see to networking, - we can take a look at the servers, - we can check operating system
health status, - and to some extent - we also understand application layer. So, this way we actually can take
a look at those log files - and configuration data and
performance data - of the whole environment. And that is the big deal. This is why this is number one on my list - why InfoSight is still the predictive
analytics tool - to beat on the market. Because it has the deepest understanding
of the whole environment. And that's a huge thing! The next thing is obviously - trying to store this tremendous
amount of data - that is collecting all the time from the environment. And for that we have a cloud service. We have a system - running somewhere in the cloud. And this is obviously the InfoSight. So, all the telemetry data that we are collecting
from the whole environment - we are reporting every now and then to
InfoSight. And this data is stored in a database - which is rumored to be the second
largest database in the world! I don't know about that but anyway it must
be large - it must be huge to store the data not only - from this one example environment - but tens of thousands of other environments around the world. different sized environments, small
and big environments- and this data has been reported
back to InfoSight - periodically, automatically, all the time - already from year 2010. So, 10 years already information
has been pouring in from - all different kinds of InfoSight/Nimble environments - around the world. And being the pioneer of predictive analytics - and being the first one on the market, - being the kind of innovator of this whole thing - is reason number two why I think - InfoSight is still leading the pack in 2020. But the data itself it doesn't have
any value to anybody - unless we can turn it into information. And how do we turn the data that we
have collected into information? Well, InfoSights or Nimble Storages
solution for this - is to have a dedicated team - of data scientists. And that team they call a Peak Team. A peak team of data scientists. With the help of AI and machine learning here. And of course the same data scientist team - is creating these algorithms all the time. So that in the future we can predict - this information that's pouring in
from the community... ...automatically. Let's say there's some situation
happening in the server here. Most probably before you even know
or the end user even knows - there's something wrong - the report of that problem already has been sent - to InfoSight. And the support team is already hard at work
trying to solve that problem. And they take their time solving the problem - and when they are ready with the solution... they have tracked down first of all the problem... and they are ready with the solution - then they will be sending a message back - to this environment that: "Hey, we found out - that after you updated one driver
in this one server - whatever server it might be - the network started to slow down - and that is because this driver - in the server that you updated - is not compatible with the VMware vSphere
version that you have installed. So that is the problem. So, roll back to the previous driver
version and that's that. So, obviously when there is a problem
in the environment - that is not well-known already to
InfoSight and HPE team - they have to take action the first time. But here is where the beauty starts! So what HPE and InfoSight does - is that we are actually mapping
every single environment. So, all of these environments
where we have InfoSight installed - we are creating this digital
fingerprint, let's say. Or digital signature. We are taking a look at all the hardware
components, - servers, operating systems, hypervisors, - storage boxes and network switches
and everything. And then we create a digital signature
for this kind of environment. Let's say this environment is blue hashtag. In reality probably that's something a bit more cryptic - like hash or something like that - but for simplicity's sake - let's say that this environment
is a blue hashtag environment. Now, we know that this blue hashtag
environment exists here in the database. Next step what we do is because we had a
problem in a blue hashtag environment - then we take a look at the database again - and we try to find out if there are other
blue hashtag environments. So, identical environments to this one. We find out that there are two blue hash
tags here in the database. Next, we locate them. So, that's a blue hashtag - and that's another blue hashtag. What we do next is we send them
information - That: "Hey, we know this with one other customer - that when you update your server driver - to the latest version - it is not compatible with the hypervisor version - that you have installed in your environment - so please do NOT do that - before you operate the hypervisor
to the supported level." So this way we are actually preventing - already to support tickets here. So, we are preventing these guys
from upgrading - because we know for a fact - that something terrible will happen
to your network - if you do that upgrading. So please read the release notes - and upgrade your hypervisor first
to the supported level. and this is the reason number three - why I think InfoSight is the most advanced
predictive analytic system out there. And that is because of the
previous two reasons: they have the deepest understanding of the whole environment - because they have cross-stack analytics. And they have been around
for the longest time. So, one more to go! I have been talking about Nimble Storage
so far almost exclusively - and that's only natural because
Nimble Storage was - the company who came up with the whole idea, - they developed it they shipped out
the first systems out there - and they made the first customers happy. But now after HPE acquisition - the InfoSight feature, the whole feature
set and idea of InfoSight - has been rolled out to many other systems as well. So, quite quickly after the acquisition - they rolled out the support for HP 3Par. Not long after that, support for servers. We are talking about ProLiants, Apollos... we are talking about Synergys, as well. Not long after that when we witnessed
the support for SimpliVity. Then obviously the latest ones: - we have Primera - and also of course the dHCI, - Nimble Storage dHCI. so, the number of supported devices that are - today reporting back to InfoSight - is tremendous. Of course I don't have the
exact numbers - but just trying to estimate - I think the number is somewhere along - half a million devices that ar that are at least capable of reporting
to InfoSight today. And what's so special about InfoSight - is that it has been designed from ground up - to support whatever devices that are - capable of first of all collecting
telemetry data - and log data from the environment comprehensively. And secondly, send it out to somewhere. It's not designed to support
only storage arrays - but basically anything that's capable of doing this. So, that's why it has been so... easy to integrate into all these other devices. I'm pretty sure if you asked HPE engineering - "easy" would not be the term basically - but but relatively easy, let's say. And that is reason number four why I think
InfoSight is superior - what comes to predictive analytics still 2020. Because they have the broadest understanding - of the data center environments. Not only the deepest... but the broadest - what comes to support the devices. Basically everybody else is
concentrating today - still heavily on the array level - while InfoSight can support whatever devices. So, that is definitely one of their biggest
advantages. That was it in short about InfoSight. if you liked the video remember to hit like - and also if you loved the video, subscribe - And hit that bell button! See you with the next one!
