Hortonworks Hybrid Cloud - Putting you back in control of your data
Immuta Overview - February 2016
1.
2. THE PROBLEM
Massive investments are being made in big data, but organizations
are stuck. Data is spread out across silos, locked down by security
policies and data owners, and data scientists spend most of their
time hunting for data, not uncovering opportunities. Frustrated by
these barriers to successful data science, we built a way to make it
easier.
3. SECURE YOUR DATA
Control your data at the most granular
level, tweak controls dynamically on the
fly, and track absolutely everything that’s
done with your data
ACCESS YOUR DATA
Access all the disparate data across your
organization from a single catalog. No more
hunting for data or writing code to connect
to various data sources.
UNLEASH YOUR DATA
Put your algorithms into production
easily, where they’ll collect and generate
even more insights. No re-coding,
security, or IT effort involved.
One product that powers your
entire data science workflow.
5. AUDITING
You can audit all actions by users against your data: who
accessed it, when, and why. You can even know what new
data (reports, analytics…) were generated from that data.
This is great not only for compliance, but for strategy: you
understand what data you have, and how your using it.
VIRTUAL REFERENCING
At the heart of Immuta’s security is the fact that we
reference all of your data virtually. You don’t have to
consolidate your data into a single, murky data lake.
And you never have to export your data to share with
others.
GRANULAR SECURITY
Because your data is not forced into a data lake, you
can apply security policies to each piece of data,
rather than a generic security policy for the whole
lake. Each piece of data is treated uniquely, and
carries its security, rules, and history with it.
DYNAMIC CONTROL
You can change policies on the fly without having to
re-tag your data. Any change you make will instantly
impact how users view and use your data.
SECURE YOUR DATA
Policies run rampant across organizations, and
missteps can be grave. Immuta’s security is so
granular and air-tight that even the most stringent
organizations can confidently let their data scientists
work.
6. ACCESS YOUR DATA
To work their craft, data scientists need access to data,
sometimes in its rawest form. But getting access to this
data is a massive challenge. With Immuta, you can
discover and access all the disparate data across your
organization from a single, secure catalog.
SECURITY IS BUILT IN
When you access data in Immuta, your view of the data is
based on the policy logic attached to that data. Write
algorithms knowing no matter what you do, you will not
introduce a security leak.
SHOP FOR DATA, VIRTUALLY
Immuta references all the data across your organization
virtually and securely, and surfaces it in a single, well-
catalogued library. Data owners expose their data
sources, you “shop” for the data you need, and then
request access. You can check out data virtually, and
check in new insights about one or many datasets.
DATA WRANGLE ONCE
Data wrangling is time consuming; share your work.
Because Immuta lets you contribute back new derived
data sources, nothing creative done with data is ever lost,
and no user will ever has to repeat a data wrangling task
ALL YOUR DATA IN ONE PLACE
As soon as you’re given access to data, it appears in
your Immuta virtual file system. You can write your
algorithms in whatever language you want, and share
and execute your code in this file system without
having to re-write anything.
7. REACT TO EVENTS IN REAL TIME
Take your successful algorithms and promote them to
“the factory,” where they will react to streaming IoT data
and produce live insights. Because Immuta’s engine
reacts to the data as it arrives, your resources are
efficiently provisioned for the workload of the moment.
No need to build scalable logic or understand elastic
infrastructure.
INHERITED SECURITY
The engine receives only the events that it has
access to, based on the security policies specified by
the data owner.
LAB TO FACTORY
What you create in the lab runs without change in the
factory. No need to change your work just to make it
production worthy; it just works. You can use your
favorite language rather than a nascent API that may
change tomorrow.
UNLEASH YOUR DATA
PowerPoint is where models go to die. Most models don’t
make it into production because of the effort involved in
programming for scale, data migration, and security. Immuta
lets you put your algorithms into production effortlessly, where
they’ll generate live insights.
12. WORKFLOW
• Drone flies over the rail line daily and
captures images
• Images streamed back for analysis
• Analysis reveals issues
• Push notification to proper systems
• Mechanic teams in the area alerted
• Automate everything!
Case Study 2
Automating rail safety with real-
time analytics and drones.
• LEGACY ALGORITHMS
• DATA SILOS
• REAL-TIME DATA
• BI VISUALIZATION