2023 XMPro Product Roadmap - Webinar
Last updated
Was this helpful?
Last updated
Was this helpful?
Join XMPro Development Manager, Daniel King as he takes you through our highly anticipated product roadmap covering the Now, Next, and Later product releases.
welcome to the XM Pro product roadmap
webinar for 2023 I'm Daniel King
development manager at XM Pro I'm based
in Sunshine Coast Queensland Australia
and I'd like to thank you all for taking
the time to attend today
I have a lot to cover off today so if
you have any questions please send them
through and I'll try to answer them at
the end
2023 has been an amazing year so far and
I can't believe it's halfway already
I won't be recapping recently delivered
items in this session but I'd like to
encourage you to go to our website and
check out our latest release notes
there are three key achievements I would
like to call out though 83 of our agency
connectors have been contributed to so
far this year which is massive I'll be
partnered with Dell Technologies to
create a validated design for the
manufacturing Edge and work with Nvidia
to validate XM Pro as GPU enabled or
accelerated and Enterprise ready
I'd like to take a second here to thank
the XM pro team for their contributions
and making all this happen I know a few
of them are watching today so thank you
the intelligent digital twins framework
is a strategy for our product
Peter Van scope fake is CEO of XM Pro
and co-chair for the natural resources
of working group at the digital twin
Consortium
here goes into much more detail on our
strategy and if you haven't already
watched this previous webinar I'd highly
recommend you do so you can find it on
our XM Pro channel in YouTube
the i3c framework is a longer term view
of where we're heading
and the product roadmap represents the
items we're working on now and next to
get us there I won't be able to cover
everything on a roadmap in 30 minutes
but I'll try my best to cover as much as
I can today
four pillars of the XM Pro product are
how we execute on the i3c framework
together they support each of our three
themes faster time to Value distributed
intelligence and secure cost deployments
the triangle layout demonstrates how
each of these Builders support each
other
and the pillars have varying impact on
people processing Technology based on
their proximity to these little labels
here
the robot items will be grouped by
pillar and I'll be talking through each
roadmap item for each of these pillars
distributed intelligence is achieved
through Ai and Engineering where people
and Technology meet to innovate execute
and augment with AI
I think that's pretty cool
we're going to start today with AI and
Engineering pillar
47 of all iot applications will be AI
infused in 2027. that's a 30 increase in
under four years that's that's that's
amazing we want to enable our customers
to be part of that 47 percent
and we're doing this with a number of AI
roadmap items
so we're going to focus on Innovation
with AI first
AI is fast growing and disrupting
markets around us and it offers new
tools to do things differently but these
tools are rapidly changing this is why
we think Innovation with AI is critical
X and pro notebooks is a new product
we'll be releasing as part of our AI
offering it supports Jupiter notebooks
allowing you to use a well-established
product without having to retrain people
or refactor existing notebooks XM Pro
notebooks can be used in data science
scientific Computing and machine
learning and data scientists analysts
and Engineers will be able to access
data to innovate within the XM Pro Suite
each user will have their own dedicated
notebook actively using the product this
makes it very scalable and cost
effective as you experiment
one of the biggest challenges for data
scientists is the access to good data to
innovate with
we're adding integration to our data
stream data so excellent Pro notebooks
can access the rich and near real-time
data in transit
this scenario isn't just limited
real-time data though in data streams
you can also run simulations on your
data streams and access that same data
from Maximum code notebooks the same way
this approach also allows users to focus
on innovation without first having to
develop software and Integrations up
front
it also allows for continuous innovation
and experimentation over time
we currently support the ability to run
simulations on streams visualize the
results in apps
front running simulations is the ability
to run a simulated event just before the
real event to predict what that result
would be
and this example we're simulating an
event right before the real event to see
what the outcome would be and in this
case it would have been a bad outcome
this allows you to build functionality
into a stream to modify the path
an event takes ultimately ensuring that
you get a good outcome
so let's talk about when these things
are coming so the now Title Here
represents an item that's actively in
development or scheduled for release
soon and the next title represents work
with developers not yet commenced or is
blocked by another item
we practice hypothesis driven
development at xmpros this means that
items can shift or change the scope
based on what we learned
development for the X and pro notebooks
has completed and the team are working
on productizing it for our supported
platforms and
Cloud providers
it will be the first available on our
free trial offering which is probably
only days away at this stage or really
close and it'll become available
incrementally on our installers for
different clouds and platforms
next we'll be releasing support to
consume stream data in those notebooks
and we're in the early stages of
planning for this
front running simulations is also in the
early stages of planning
once you've innovated and produced new
AI Solutions you need a way to execute
them
running AI in our streams is one way to
do this and up until recently you'll be
limited running these on the CPU
the GPU is well suited for running
certain workloads and involve a lot of
concurrent processing
moving processing to the GPU can vastly
improve how much processing you can do
in the same time frame compared to the
CPU mathematical algorithms Machine
Vision neural networks and deep learning
are examples of workloads that involve a
lot of concurrent processing
in this example we're using a Machine
Vision to process apples and oranges
differently in our data stream
images of our fruit are ingested by a
data stream as an agent compares the
images pixel by pixel to determine if
the if it's an apple or an orange
by using the GPU to run pixel comparison
algorithms we can process a lot more
images
this stream can still be run on the edge
Fogle cloud
and it can also dynamically detect if
there's a compatible Nvidia GPU and use
it falling back to the CPU when the GPU
isn't available
there are a handful of scripting
languages that are fast becoming
standard for executing AO workloads by
supporting these scripting languages we
support customers in running their
existing Solutions in our streams this
reduces the need to refactor or retrain
staff and you can adopt new
functionality faster as is the large
open source Community currently
innovating using these scripting
languages
we're extending our current agents to
support more of these scripting
languages
we're also adding support for more
governments around these scripts in this
example you can see we can run a script
directly in an agent or reference the
latest published version of a script
from the common storage layer
for now I think of the common storage
layers kind of like GitHub and I'll talk
about this later
these scripts can be edited from the X
and pro notebooks allowing users to
operate completely in our suite of
products
data streams are executed for us to buy
our stream hosts and these stream hosts
can be run on the edge of fog and Cloud
allowing you to distribute your
processing fully strict scripts uh
providing like a very flexible and
composable architecture
thank you now we've recognize you may
have your own machine learning
operations or mlops
they're focused on streamlining the
process of taking machine learning
models to product and then maintaining
and monitoring them
mlop space it's it's growing very
quickly and there's a vast variety of
different third-party Solutions our
approach here is to create agents to
integrate these third-party Solutions
mlflow is our open source platform for
managing the end-to-end machine learning
life cycle
and is fast becoming the the most
popular in the space
we have created an ml flow agent that
can get the latest published version of
a model
and invoke it with data from the our
data stream and then return those
results back this prevents the the need
for a data science to actually go into a
data stream and update that stream each
time a new version of the model is
created
it also allows heavy processing of ml
models to be run on the customer systems
optimize already for these workloads
it also allows access to be restricted
to the data stream supporting the
principle of least privilege
mlops can be orchestrated from XM Pro
notebooks allowing you to run
experiments create models publish those
models and then invoke them directly
from The Notebook then using the results
foreign
GPU acceleration support is currently
available you can design Aid install
leverage Nvidia gpus using agent
documentation and we'll be extending
this documentation to provide more
examples and support
we're also working on adding GPU
acceleration support to some of our
existing agents
we currently have a python script and
we're close to releasing an R script
agent we'll be adding more governance
controls and integration to these to
these scripts using the common storage
layer after it's released and I'll talk
about talk more about that common source
layer layer later
we've recently developed the ml for
agent for integrating to customer lmops
and we're updating our public docs right
now before we actually release it
now the innovating and executing with AI
there's a great opportunity to augment
what we have now with these results
we currently have recommendations
capability and for those that aren't
familiar at a very high level you can
create recommendations with rules that
when are met generate alerts
we found that existing recommendations
can be augmented with AI anomalies can
be dynamically detected and visualized
with inside the recommendations
new recommendations can be automatically
generated or discovered
and rules within those recommendations
can be created or augmented
and alerts themselves can be augmented
with copilot like functionality and
assisting assisting engineers in in
resolving these alerts
chat GTP and open API services are
continuing to provide new and easier
innovative ways of doing things
accent Pro notebooks will be released
with a sample notebook demonstrating
integration to chat GTP in this sample
notebook it includes an end-to-end beer
quality example which incorporates chat
GTP and who doesn't like quality beer
data stream designer Integrations will
allow you to augment your event data
with chat GTP and app designer
Integrations will allow you to make
custom calls
from your apps enabling co-pilot like
functionality
the security of our data we will we uh
security fuel data and and how we
implement this is really important to us
we use the third-party services like GTP
takes a lot of our consideration and
we'll be factoring this into you know
how we solve these problems and deliver
these new product roadmap items
visualization of AI outputs is a really
powerful and effective way of
communicating the insights you're making
excellent Pro notebooks will allow you
to generate visualizations that you can
share in apps in this example we'll be
using chuckttp to generate a linear
regression visualization for us
it's using the data for my data streams
and it's checking to be savings a lot of
time because we're not having to find
the suitable libraries and write the
code to do this ourselves
Excel Pro notebooks will be released
with chat GTB integration support and
samples on how to get going quickly you
can currently integrate you integrate to
check GTP via our Python scripts and run
these in agents right now
and these Python scripts can be designed
and managed in X and pro notebooks
we will have importable end-to-end
Solutions as our starting point to get
you going faster for our common use
cases and additionally planning for app
designer and data stream designer
integration
as this is currently underway and it
will provide more additional no code
approaches
AI generated or discovered
recommendations is in the early stages
of planning there are some core
technical pieces that are needed before
we can start this work and the
development team are busy work on these
right now
generating analytics and visualizations
is currently available through a mix of
approaches we're currently planning on
how to extend these approaches to closer
integrate x and pro notebooks allowing
from a more seamless experiment
experience from right through from
Innovation through to augmentation and
the visualization
all these new AI roadmap items need
distributed computing and network
infrastructure management to support an
edge ecosystem
this is the cloud to Edge continuum
we currently support deployments to a
wide variety of cloud providers and
on-prem platforms using cloud and
platform native services
this creates additional overhead for us
and our customers as we add more
features demanding more capabilities in
this area
these environments can be configured in
many different ways and also change and
sometimes without our control and this
makes it difficult for us to deploy and
support our products simply as
seamlessly
there are new more modern approaches
that are quickly become industry
standard for the type of architecture
that underpins our suite of products
these more modern approaches require us
to change our deployments to embrace
them
we'll be creating a cloud agnostic
deployment for our suite of products
that will become the default deployment
method
the aim is to provide a product that is
more portable allowing for business
better business continuity planning more
reliable with self-healing properties
and more performant allowing for better
performance for Less cost
now we currently support distributed
deployment and in this example we have
products used most by our users in the
cloud in regions closest to those users
making the products feel fast and snappy
in the fog we have our AI capabilities
closest to where most of our data is
reducing the time to transmit large
volumes of data over the wire and on the
edge we have our stream hosts these are
our engines that run our data streams by
putting them on the edge we put them
closest with the devices and Edge
systems are
the challenge with this approach is
managing all these Edge deployments
without some form of centralized
monitoring alerting and management it
can become difficult and time consuming
for iot to manage these systems
this is why we're building support for
third-party systems that provides
centralized monitoring alerting and
management capabilities
we're adopting industry standards and
best practices to allow us to easily add
additional providers over time and this
also provides customers with the
flexibility to use their existing
providers if they have them
the solution would be composable with
the ability to select different
providers from different functions
instead of having to commit to one
system to do it all
we're bringing devsecops capabilities to
our sweeter products
we're not building these capabilities
into our products but again leveraging
existing industry standards and best
practices to allow you to use your own
providers
in this example we have a pre-pro
environment that we want to design our
streams and apps before we publish them
to production
these new versions of apps and streams
are automatically exported to the common
storage layer and from there UI test
automation can be run leveraging our XM
Pro test automation Library
secure testing can be performed on any
changes to say Integrations to mitigate
any potential vulnerabilities and when
all these pre-checks pass the change is
then approved and these versions can be
published to our products or your
product environment or the customer
support environment
we can take a similar approach to
upgrades to X and pro new versions can
be automatically and programmatically
installed in pre-prod manual regression
or just automated UI testing performed
and then deployed to production
we've just released automated database
installs and upgrades
which allows for a fully automated
deployment our current deployment
options required a customer though to
set up Automation in their third-party
systems
best practices have evolved since some
parts of the process were first
developed and we need to redesign these
and we'll provide be providing improved
deployment automation incrementally
product by product
we've completed several proof of
Concepts around Cloud agnostic
deployment
with commence development and are taking
our product by product approach and
we'll continue to support cloud data for
some time shifting to Cloud agnostic
deployments as default and they're done
we're looking to cater for customers who
want an out-of-the-box experiments where
they can use all the default deployment
configuration that we provide right
through to those that want to consume
the individual product artifacts and
build them into their own custom
deployment
we've completed a proof of concept I
distributed deployment management based
on our Cloud agnostic approach and we're
in the planning state of considering a
mix of internal and external third-party
tooling to be rolled out incrementally
faster time to value is achieved through
roadmap items that combine people
and process to accelerate transformation
we're providing more support for what
you have when it comes to visualizations
we use blocks to compose apps and
already have a large library of blocks
to do this however some scenarios it's
quite difficult to create a block for
example
if a customer system uses an
incompatible technology or uses an old
technology that performs slowly or is
just custom or an in-house system
meta blocks will allow us or you to
create custom blocks to overcome these
challenges
it'll allow you to run visualizations
side by side that traditionally just
wouldn't be possible
with this capability you will be able to
create an industrial metaverse that runs
on the same backbone of event data
without first needing to re-platform
your existing exist existing systems to
like a common technology
this gives you the flexibility to delay
deciding when that common technology
needs to be and what it needs to be
until that Industrial metaverse
Technologies and processes stabilize and
you can make a better decision about
that
I've mentioned the common storage layer
a number of times already in this
presentation
and that's because it will unite
artifacts across XM Pro think of it as
kind of like GitHub
except we'll be implementing it in a way
that allow you to choose the provider so
it doesn't have to be GitHub it'll also
be comprised of several different
Technologies allowing for the best
technology for the job and using best
practices to do it
the common storage layout will allow you
to collaborate faster and safer
providing well-known interfaces and
governance controls to do this
anything you can currently export or
import can be managed here and all the
new artifacts discussed today will be
stored here
the provider you choose for the common
storage layer will also support
integration most likely and allowing you
to automate and tie in artifacts from
Max and pro into your current existing
systems
we're introducing some new artifacts the
common series later get you started
quicker and keep you innovating
blueprints are pre-built Solutions
providing end-to-end working samples
accelerators can be imported into your
environment as a starting point
from which you then can extend from
and patents are pre-configured data
streams at servers building blocks that
you can compose and extend out
you'll also be able to create your own
versions and collaborate internally and
externally
so we currently have a library
predefined Solutions in our XM Pro
GitHub project which you can now check
out
we'll also continue to add to these over
time and next we're looking at adding a
library UI into our accent pro products
to access and integrate them a little
bit more easier
we've completed a perfect concept of
meta blocks and are planning on studying
development soon and we've selected
several new and existing blocks to be
candidates for that first release
we're currently using several providers
for my common storage layer internally
for collaboration and we'll be exposing
these publicly
we're beginning planning We Begin
planning on how we'll integrate these
products in the UI shortly
where process and Technology meet we
need strong governance to be secure
across deployments and we do this with
zero trust architecture
zero trust architecture is not new for
us at X and pro and something we take
really seriously
we aim to give you more control over
what users can access with finer grained
Access Control
the product rights will align closer to
API endpoint functionality and not UI
components we'll still have product
roles to manage any additional
complexity this might introduce and in
this diagram you can see a user
accessing data via connector or
recommendations functionality previously
these would have been these five rights
here would have been three different
rights this change allows for greater
control over what users can access
and by aligning these rights to API
endpoints it makes it simpler to apply
the principle of least privilege at
scale
we currently run several security scans
every three months months and uh and
publish these results to our website
in addition to this we'll be running
incremental scans on every code change
for each product to pick up potential
vulnerabilities earlier this change will
ensure that any new vulnerabilities are
identified and addressed as soon as
possible
so XM Pro currently supports integration
to Azure ad ID provider providing an
easy way to leverage existing user
logins and security policies and
features like SSO and MFA
we're extending this functionality to
include support for more scenarios like
a Federated tenancy model
this integration has been done using
industry standards making it easier for
us to add new providers when it's
necessary
as we implement the edge to Cloud
Continuum and monitoring and managing
product performance is critical
we need it we're implementing support
for application Performance Management
tooling and we're doing this using
industry standards so adding new
providers will be easier using this
integration you'll be able to monitor
performance of our products
be alerted when there's issues and if
needed do basic troubleshooting yourself
so we're applying finer grade access
controls incrementally to each of our
products with the first product in
development now
we have continuous security scanning set
up on most of our products are in the
process of updating the remaining right
now
and we currently have support for Azure
ID and like I mentioned before we're
just extending this to support more
scenarios like a Federated tenancy model
and we're currently adding support for
Azure app insights as a APN provider
so
we'll be running webinars on the items
shown today as they become available and
diving into a lot more deeper detail
if you have any questions please send
them through and I'll try to answer them
now thanks
thanks Dan um so we've got two questions
here already
um the first one is where can we find
more information on XM Pro AI
okay great question ah thanks Sarah
so that segues really nicely uh next
month Gavin Green will be doing a
webinar on AI and he'll go into a lot
more detail
um I'd encourage you to sign up for that
and attend
um we're very close to releasing X and
pro notebooks in our free trial soon
um I think it's a couple of days away at
the stage just keep an eye out for that
uh sign up have a play around with our
free trial and the different AR
capabilities that are in there and I
also believe we're publishing a web page
specifically around Ai and our website
and the features that you can leverage
there so keep an eye out for that as
well
awesome there's another question here um
it says with Edge to Cloud continuum
providing any
s around containerized architecture
good question uh yes really really
excited about this one
um so actually we currently have Docker
images for our stream hosts these are
the engines that run our data streams
we're adding Docker images for other
products and we'll be creating a
container registry that will be
publishing these two
um we will likely do webinars on this as
this becomes available
um
so yes
awesome and then one very important
question asking will we receive a copy
of the webinar from today
yes yeah this is probably where I do the
uh like And subscribe Us on YouTube
um yes we'll be putting this on YouTube
on our XM Pro Channel all of our
webinars are on there including uh Peter
Scott fakes previous one I mentioned
please jump on there and look at it look
at it um Sarah do you know how long
it'll take for it to get up there
um hopefully we'll get it up today
sometime later on today okay
thanks
that's it for the questions
oh I head over to you oh thanks
okay
um so if you just want to change science
then
cool
thanks everyone um so thanks very much
Dan for taking us through and thanks
everyone for joining us if you're
looking for more information uh you can
contact Dan directly or the team via
these email addresses as Dan mentions
we're running these webinars monthly and
our next session will be on XM Pro AI
presented by Gavin Green so you can
register by scanning the QR code there's
a link in the chat box and I'll also be
sending the link
um shortly when we send the recording
out sometime later on today and we look
forward to seeing you all next month
thank you very much for joining
thank you