Are Data Lakes for Business Users?

Hosted by DM Radio

Eric Kavanagh

CEO

Steve Wooledge

VP of Marketing

Wayne Eckerson

Founder and Principal Consultant

Data lakes took root because they provided an instant sandbox for data scientists to explore raw data sourced from operational, analytical, and external systems.

As data lakes have matured, organizations have begun to ask whether they can be used to support more traditional business users — executives, managers, and front-line workers who want to learn from curated data and dashboards, not wrangle with raw data and SQL.

Please join us for this joint webinar with the Eckerson Group and the Bloor Group. In this webinar, several industry experts will discuss:

  • The evolution of data lakes and analytical tools
  • Whether business users are really taking advantage of these new data constructs
  • How organizations can measure the efficacy of their data lakes with seven key metrics

More About the Presenters:

Eric Kavanagh

Eric has more than 20 years of experience as a career journalist with a keen focus on enterprise technologies. His mission is to help people leverage the power of software, methodologies and politics in order to get things done.

Steve Wooledge

Steve is a 15-year veteran of enterprise software in both large public companies and early-stage start-ups and has a passion for bringing innovative technology to market.

Wayne Eckerson

Wayne is a long-time thought leader in the BI and analytics field who has a passion for helping business and technical executives and managers strengthen their leadership skills and increase their effectiveness to drive positive change in their organizations.

Transcript

[00:00:00.000]
welcome my name is Shannon

[00:00:02.100]
camping on the cheap digital manager at a diversity

[00:00:04.299]
would like to thank you for joining today's DM

[00:00:06.500]
radio Arcata legs for business users

[00:00:08.500]
sponsored by Arcadia data continuing

[00:00:12.400]
conversation from a live game radio broadcast

[00:00:14.400]
a few weeks ago which if you missed you can

[00:00:16.600]
listen to it on demand DM

[00:00:19.000]
radio. Biz under podcast

[00:00:21.100]
just a couple of points to get us started number

[00:00:24.199]
of people that attend the sessions you will be muted

[00:00:26.399]
during the webinar if you'd like to chat with

[00:00:28.500]
out there with each other we certainly encourage you to

[00:00:30.500]
do so just click the chat icon in the upper-right hand

[00:00:32.600]
corner for that feature to

[00:00:36.100]
an infection in the bottom right-hand corner of your screen

[00:00:38.200]
or if you like sweet we

[00:00:40.299]
encourage you to show how it's a question sites what are

[00:00:42.299]
using hashtag DM radio as

[00:00:44.700]
always we will send a follow-up email within 2

[00:00:46.700]
business days containing lyrics to the fly's the

[00:00:48.899]
recording of the session and additional

[00:00:51.100]
information requested throughout the webinar

[00:00:59.899]
oh and welcome welcome

[00:01:03.100]
everybody thank you Shannon yes indeed

[00:01:05.299]
it's time for another guy

[00:01:12.400]
right in our data Lakes for business users

[00:01:14.700]
obviously that is a slightly rhetorical

[00:01:17.200]
question the apparent gold

[00:01:19.299]
diggers

[00:01:25.000]
today Steve Willets of Arcadia data

[00:01:27.099]
of course you are truly in the middle there and my good buddy

[00:01:29.299]
Wayne Atkinson of X

[00:01:31.400]
and group and I go way way back

[00:01:33.500]
to the old history

[00:01:36.200]
together focused on things like state

[00:01:38.400]
of warehousing now courses data and lakes

[00:01:40.700]
and is a concept I wanted to touch on Drake

[00:01:42.799]
quickly before I hand it over to Wayne to give

[00:01:45.099]
me some results from his assessments

[00:01:47.200]
and that is the whole concept of data science

[00:01:49.900]
we keep hearing about data science

[00:01:51.900]
cuz it's one of my favorite quotes from

[00:01:54.000]
the movie Nacho Libre esqueleto

[00:01:56.000]
claims

[00:02:00.099]
right we hear all about signs these

[00:02:02.099]
days in data science as well we

[00:02:04.299]
all know the numbers don't lie but they sure can

[00:02:06.400]
be misuse or misrepresented Topic

[00:02:13.000]
in perspective on what we're trying to accomplish here

[00:02:15.199]
today what we're trying to accomplish in

[00:02:17.199]
the broader business intelligence analytics

[00:02:19.800]
big data market we're trying to use

[00:02:21.900]
data to get insights to

[00:02:24.000]
be able to make better decisions for our business

[00:02:26.300]
the mission is the saying the mission has not

[00:02:28.300]
changed the tools have gotten much more

[00:02:30.400]
powerful we not talk about dado blades

[00:02:32.599]
as opposed to data warehouses and they

[00:02:34.599]
are in fact very different things are

[00:02:36.900]
designed very differently they were developed

[00:02:38.900]
in different eras of this

[00:02:40.900]
industry and let's face it they

[00:02:42.900]
were a full set of constraints many

[00:02:45.400]
many years ago Wednesday to warehouses

[00:02:47.400]
were designed around which they were

[00:02:49.400]
built processors were slow pipes

[00:02:52.000]
worth in for example memories

[00:02:54.199]
I was expensive so

[00:02:56.199]
all these factors really dictated

[00:02:58.400]
what had to happen in creating

[00:03:01.000]
did warehouses are also

[00:03:03.000]
extremely expensive it's a good

[00:03:05.099]
pair of data warehouse appointment today versus

[00:03:07.199]
125 years ago it's

[00:03:09.199]
astonishment the price difference is gone

[00:03:11.300]
from millions of dollars to hundreds

[00:03:13.400]
of thousands of dollars to even $100,000

[00:03:16.199]
or less depending upon to use case

[00:03:18.500]
in the complexity of communication

[00:03:21.900]
is hard to do I think people

[00:03:23.900]
take medication for granted and

[00:03:26.099]
at the end of the day if you're not communicating clearly

[00:03:28.500]
with your team with your business users what

[00:03:31.000]
is it to learn what you can glean from the data

[00:03:33.199]
been really you've been on a Fool's

[00:03:35.199]
errand until it's something to be considered

[00:03:37.400]
here so science

[00:03:39.400]
data science I think it is applicable these

[00:03:41.800]
days I think that is an accurate term that we can

[00:03:43.900]
use to describe some

[00:03:46.000]
of the more robust well-thought-out

[00:03:48.199]
and efficient environment for

[00:03:50.400]
managing status but I think

[00:03:52.500]
there's a significant disconnected our culture

[00:03:54.800]
today when we think of this term

[00:03:56.900]
is science I think a lot of people believe that

[00:04:00.199]
represents a virtually installable

[00:04:02.800]
version or representation

[00:04:05.400]
of reality and that's just not true at

[00:04:07.500]
all scientific discipline

[00:04:09.500]
and it relies on a methodology

[00:04:11.800]
aka the scientific method

[00:04:14.000]
which is applied appropriately

[00:04:16.500]
and effectively and efficiently and responsibly

[00:04:18.699]
can give us great insights about

[00:04:20.699]
the world around us but remember

[00:04:22.899]
axiomatic to the

[00:04:25.000]
scientific method fundamental intrinsic

[00:04:27.399]
to the scientific method is a commitment

[00:04:29.800]
to Forever questions

[00:04:31.800]
your data your processes

[00:04:34.100]
your hypotheses and even

[00:04:36.300]
your conclusions so again.

[00:04:38.600]
My point is that we need to take the term

[00:04:40.600]
science or the pit of a grain of salt here and

[00:04:43.000]
find this change their mind being

[00:04:47.699]
bad for your member house it's really

[00:04:49.800]
scared of Peter kitch about 10 or 15 years ago

[00:04:52.000]
all the cholesterol in eggs you're

[00:04:54.000]
going to get a heart attack and then what happened

[00:04:56.199]
and it came out with good cholesterol and bad cholesterol right

[00:04:59.000]
okay

[00:05:00.199]
what does that mean I think

[00:05:02.300]
the point is it's fine that will change their

[00:05:04.300]
minds about things sometimes it's frankly

[00:05:06.300]
can also be paid by large organizations

[00:05:08.899]
to say things that they stay probably believe

[00:05:11.000]
to distort what

[00:05:13.300]
is the reality that were all trying to better

[00:05:15.399]
understand so it's just one

[00:05:17.399]
quick example of how much we really don't

[00:05:19.600]
know these days when will the lava

[00:05:21.600]
flow stop in Hawaii we

[00:05:24.199]
just don't know and the reason we

[00:05:26.199]
don't know it's because the Earth is really large

[00:05:28.399]
environment and things like

[00:05:30.399]
volcanoes are extremely hard to protect

[00:05:32.500]
their very powerful very complex

[00:05:35.000]
and we just don't fully understand what's

[00:05:37.100]
going on does the magnitude

[00:05:39.100]
of those the problem space if

[00:05:41.100]
you will trying to understand where the hell I was

[00:05:43.100]
going to blow where the pictures will come from

[00:05:45.199]
next what that volcano may do

[00:05:47.199]
next week just don't know and

[00:05:49.300]
so I think it's just paid to remember that

[00:05:52.199]
data will always require analysis

[00:05:54.899]
no matter how efficient you are with a data

[00:05:57.000]
Lake management project for example

[00:05:59.100]
you're still going to analyze

[00:06:01.100]
that dated to put it into contacts to

[00:06:03.600]
view it in reference to

[00:06:05.699]
your current situation than

[00:06:07.800]
the historical data that you may have no

[00:06:10.399]
data is ever going to give you the complete and total

[00:06:12.399]
answer of the story because you have to

[00:06:14.500]
come up with a story yourself so

[00:06:16.800]
leveraging what you know is important to take

[00:06:18.800]
Sammy takes boxing analytics

[00:06:20.899]
and Big Data are all very useful invaluable

[00:06:23.300]
if we understand what we know

[00:06:25.300]
I know roughly what we're

[00:06:27.300]
doing and so with that I'm going to hand it off to

[00:06:29.399]
my good buddy Wayne Eckerson was

[00:06:31.800]
going to talk about the assessment

[00:06:33.899]
that we've done on behalf of our katydid

[00:06:36.000]
in our end users about data lakes and

[00:06:38.100]
the value that it provides for business users so

[00:06:40.399]
with that way never seen I hand it over to you thank

[00:06:43.699]
you Eric it's great to be here

[00:06:45.699]
with you once again and date

[00:06:47.899]
of birth city and everyone in the in the audience

[00:06:51.600]
whoa that

[00:06:54.199]
didn't work that

[00:06:57.699]
image for some reason it's

[00:07:00.600]
not showing up that image

[00:07:02.600]
of a data Lake and

[00:07:05.399]
the

[00:07:15.699]
data Lake yeah

[00:07:18.199]
I saw this this webcast

[00:07:20.199]
it's about the business value of data

[00:07:22.500]
lakes and it's Eric mentioned

[00:07:24.800]
date Alexa Rose

[00:07:26.899]
number of years ago almost

[00:07:29.300]
10 years ago when Cloudera was founded

[00:07:31.399]
really to address a

[00:07:33.399]
lot of frustration on the business side with

[00:07:36.500]
the data warehousing you

[00:07:38.899]
know too slow too

[00:07:41.000]
hard to the design

[00:07:43.000]
too hard to change very costly scalability

[00:07:47.000]
Tech capped out in a couple terabytes

[00:07:49.300]
really didn't really handle

[00:07:51.600]
unstructured data very well so

[00:07:54.500]
I'm fast forward

[00:07:56.500]
to the day awake and Hadoop and

[00:07:58.600]
that made a lot of people

[00:08:00.800]
happy however

[00:08:02.800]
it did not replace the day to Warehouse

[00:08:05.199]
what did they delay became

[00:08:07.899]
in essence very quickly

[00:08:10.000]
was not a day

[00:08:12.000]
to Warehouse replacement but really an

[00:08:14.399]
ultimate ideal sandbox

[00:08:16.399]
for data scientist

[00:08:17.699]
or power users really

[00:08:20.100]
wanted what they've always wanted

[00:08:22.100]
historically is big giant

[00:08:24.300]
data dumps and then

[00:08:26.500]
to get icy out of the way and it

[00:08:28.500]
did like the sensory was bad just

[00:08:30.600]
put all the data in one place and

[00:08:32.899]
then let me go in and navigated

[00:08:35.000]
manipulated manage it

[00:08:37.100]
and analyze it

[00:08:39.100]
and create models from

[00:08:41.200]
it so the data Way Grill

[00:08:43.500]
in its first Incarnation turned

[00:08:45.700]
out to be a great for data scientists

[00:08:47.700]
and data analyst

[00:08:49.799]
how are users he wanted access to the

[00:08:51.899]
raw data it

[00:08:53.899]
really wasn't a replacement for the warehouse

[00:08:56.500]
that supported standard

[00:08:58.899]
dashboards and reports for

[00:09:02.000]
I would call Casual users

[00:09:04.299]
executive Managers from my workers

[00:09:06.399]
really needed Taylor and accessed

[00:09:08.500]
information so

[00:09:13.200]
yeah we've we

[00:09:15.200]
passed recently when we got together at

[00:09:17.299]
Arcadia what about State

[00:09:19.799]
wakes and regular users who don't

[00:09:22.700]
know sequel python or Java which

[00:09:24.700]
was a tools of choice for

[00:09:26.700]
a dupe type of processing

[00:09:29.600]
in analytics need a

[00:09:31.600]
graphical interface to analyze

[00:09:33.600]
data also known as bi

[00:09:36.000]
tool cool require

[00:09:38.200]
clean curated aggregated

[00:09:40.200]
data another would someone typically

[00:09:43.000]
nighty to go in and

[00:09:45.000]
take the raw data and then nip

[00:09:47.500]
you waited so clean

[00:09:49.600]
it is integrated so that deep casual

[00:09:51.700]
users can make sense of it without

[00:09:53.700]
having to do all that manipulation ourselves

[00:09:56.100]
and who needs sub 2nd

[00:09:58.299]
performance Perry performance and

[00:10:01.200]
use a data in reports and

[00:10:03.399]
dashboards that are highly Taylor to meet

[00:10:05.399]
so they only see what they need and nothing that

[00:10:07.399]
they don't try

[00:10:09.500]
to find your past that really meet

[00:10:11.799]
their their needs to glance

[00:10:14.200]
at KP eyes and

[00:10:16.299]
and take action if

[00:10:18.299]
things or arrive so

[00:10:20.600]
it's

[00:10:24.700]
nothing a good thing for

[00:10:26.799]
regular users regular Joes

[00:10:28.899]
if he will executive

[00:10:31.399]
managers front-line workers even

[00:10:33.399]
customers and suppliers but

[00:10:35.899]
we got to go to the car came inside let's

[00:10:38.000]
let's test this let's do an assessment

[00:10:41.200]
and figure out if

[00:10:43.200]
this is still the case if

[00:10:45.299]
they do legs are still just for power users

[00:10:47.299]
are not so

[00:10:49.299]
we did an assessment came up

[00:10:51.299]
with the survey of 22 questions

[00:10:53.500]
for took about 5 minutes to complete once

[00:10:56.799]
they completed it ever seen group assessment

[00:10:59.399]
or surveys generator Dynamic

[00:11:01.700]
support as you can see here on the right Ira

[00:11:04.700]
personalized it's it gives him a

[00:11:06.700]
score comparison to everyone else

[00:11:09.100]
over all in by category

[00:11:11.399]
was conditions for next steps based

[00:11:13.600]
on on their rank in

[00:11:15.700]
the scoring so that

[00:11:17.799]
Stephen still running now and I encourage

[00:11:20.100]
you to go out and take it at the link below

[00:11:22.100]
to assess

[00:11:24.200]
the value of your day

[00:11:26.200]
like if you have one for your

[00:11:28.299]
regular business users

[00:11:30.799]
so as of

[00:11:32.799]
April 20th when I put the slides

[00:11:35.100]
together 100

[00:11:37.500]
almost 200 had started the

[00:11:39.500]
Assassin 262 at completed at

[00:11:41.600]
93

[00:11:43.700]
have a date of lake in production and those are

[00:11:45.700]
the folks that we really want to focus on I'll

[00:11:48.700]
go to 74% were from North America

[00:11:51.000]
and about half was fairly large organizations

[00:11:53.600]
with more than 10,000 police is

[00:11:59.399]
based on that subset of the respondent

[00:12:01.700]
base I think we're up to almost

[00:12:03.899]
250 respondents now we'd

[00:12:05.899]
love to have you get us up to 300

[00:12:07.899]
so I'll write out that you

[00:12:10.000]
are out and go it only takes

[00:12:12.200]
5 minutes or less to complete the assessment

[00:12:14.399]
and you get your own personal free

[00:12:17.500]
report

[00:12:20.000]
so what did

[00:12:22.000]
we find out first of all surprisingly

[00:12:25.299]
a little bit is that

[00:12:28.200]
Ford 800 x most people almost

[00:12:30.399]
two-thirds are using to do for the

[00:12:32.399]
day like and

[00:12:34.399]
I hope that's not too surprising did

[00:12:36.500]
awake it's become synonymous with a dupe

[00:12:38.500]
but in the last couple of years we've

[00:12:40.600]
seen a real rush to move the

[00:12:42.799]
state of eggs into the cloud

[00:12:45.100]
and replace a group with Cloud object

[00:12:47.399]
stores which is which are

[00:12:49.399]
currently running at 14%

[00:12:51.500]
of respondents in our

[00:12:53.700]
our pool not

[00:12:56.299]
sending 17% or running

[00:12:58.500]
their big lake in a relational database

[00:13:00.500]
and some of you might think

[00:13:02.600]
that's an anomaly but

[00:13:05.000]
truly if you used to give me method

[00:13:07.100]
of design a day to Warehouse he

[00:13:09.399]
always called for a staging area essentially

[00:13:11.799]
a place where you put your raw Data before

[00:13:14.500]
you turn it into third normal form before

[00:13:17.500]
you crazy and

[00:13:19.600]
push updated March from that

[00:13:22.299]
I also 6% said no

[00:13:24.399]
sequel database nosql

[00:13:26.500]
days is not an analytical database

[00:13:28.500]
by any means but it's only can hold a heck

[00:13:30.600]
of a lot of data that can

[00:13:32.600]
be used for analysis

[00:13:35.600]
second question here how do most users

[00:13:37.899]
query the day like in this was very surprising

[00:13:40.299]
you know David scientist

[00:13:42.399]
tend to prefer tools like python

[00:13:44.600]
Pearl Java and

[00:13:47.500]
other coating type languages

[00:13:50.000]
or in the head tube for old

[00:13:52.000]
Pig Hive tools

[00:13:54.399]
like that or just plain

[00:13:56.500]
sequel if so the

[00:13:58.600]
data in had to piss inside the

[00:14:00.799]
park a files in,

[00:14:03.299]
format so we are actually

[00:14:05.299]
pleasantly surprised to see the more than half

[00:14:07.399]
or using a point B quick

[00:14:09.500]
visual bi tool the

[00:14:11.600]
court of the Dead awake so

[00:14:13.899]
that that was surprising now

[00:14:16.600]
I will say both

[00:14:19.100]
floor group and ungroup in Arcadia

[00:14:21.299]
promoted this survey and

[00:14:24.200]
each delivered a

[00:14:26.200]
bunch of a number so

[00:14:28.399]
there may be some Buys in there but I don't

[00:14:30.500]
think too much so I think we can

[00:14:32.500]
trust it this data is generally representative

[00:14:35.000]
of the market

[00:14:38.600]
okay then

[00:14:40.600]
we asked what you

[00:14:43.100]
know where have you deployed your data like and

[00:14:45.899]
you can see here that the large percentage

[00:14:48.299]
is still on Primus public

[00:14:52.100]
cloud is ranges

[00:14:54.500]
between 19 and 20 18%

[00:14:57.000]
so less than 20% there and

[00:15:00.500]
between 15.6%

[00:15:04.100]
I have a hybrid environment both on premises

[00:15:06.299]
and Cloud.

[00:15:08.500]
I just said that we're seeing a

[00:15:10.500]
large gravitation

[00:15:12.500]
towards the clouds or running day lights

[00:15:14.500]
with this this chart actually

[00:15:16.899]
contradicts that and it shows

[00:15:19.000]
that company so I got to play dead lights

[00:15:21.000]
in the last 2 years or more likely

[00:15:23.500]
to deploy on premises so

[00:15:25.700]
I'm not sure I quite understand that that

[00:15:27.799]
kind of runs counter to what we're seeing

[00:15:29.899]
generally out there or at least anecdotally

[00:15:32.899]
but numbers don't lie is

[00:15:34.899]
that Eric would like to say so

[00:15:37.100]
I have to discuss that in a little bit

[00:15:43.100]
can business users explore

[00:15:45.100]
data to get the views they want so

[00:15:47.600]
this is the part

[00:15:50.299]
and parcel of what power

[00:15:53.299]
users always do and Casual

[00:15:55.299]
users to some extent do and

[00:15:57.399]
you can see hear that more than half almost

[00:15:59.799]
two-thirds agree

[00:16:01.899]
or strongly agree with that statement

[00:16:04.200]
so the day

[00:16:06.200]
like really is exploration area

[00:16:08.799]
of Discovery area and

[00:16:12.399]
if users are using bi tools then

[00:16:15.100]
we have to admit

[00:16:17.500]
that a large percentage of those users

[00:16:19.899]
are are casually sources do want to do exploration

[00:16:23.399]
Racine hear that the date of Lake far

[00:16:26.000]
from being a day of swamp is actually

[00:16:28.000]
providing information and data that users

[00:16:30.299]
find trustworthy and

[00:16:32.700]
enables them to make better decisions of

[00:16:35.200]
course that's the whole point of using data

[00:16:37.399]
is to improve

[00:16:39.600]
your decision-making improve

[00:16:42.000]
outcomes for the business so it's it's great

[00:16:44.200]
to see that over 50%

[00:16:47.299]
agree and 70%

[00:16:49.700]
strongly agree with that statement

[00:16:55.100]
this is

[00:16:57.100]
another surprising morning we asked about query

[00:16:59.399]
performance and 50%

[00:17:03.299]
agree or strongly agree with the statement

[00:17:05.299]
that the daylight provides consistent performance

[00:17:07.599]
when you think about it I do

[00:17:09.700]
quiz to find you some batch environment

[00:17:12.000]
and only recently has become

[00:17:14.000]
interactive with sequel

[00:17:16.200]
interface

[00:17:18.200]
so things are moving very fast

[00:17:20.299]
in the day like world and

[00:17:23.599]
Abel support out fast

[00:17:26.900]
query performance in response

[00:17:29.299]
times the next question

[00:17:31.500]
about the accuracy of analytics

[00:17:33.799]
in the day awake

[00:17:36.599]
and that's another reinforcement

[00:17:39.599]
of the notion that these day lights aren't they the swamps

[00:17:42.000]
and that people with bi tools and

[00:17:44.099]
not only make good decisions but

[00:17:46.500]
trust the day that they're working with their

[00:17:51.299]
we also did a lot of analysis

[00:17:53.299]
by company size

[00:17:55.400]
and we didn't find much variation between

[00:17:58.799]
large and small companies

[00:18:01.000]
although this chart will

[00:18:03.099]
show you that very

[00:18:05.200]
large organizations with over a hundred thousand

[00:18:07.700]
employees are a little bit more advanced

[00:18:10.400]
47% strongly

[00:18:12.400]
agree that isn't

[00:18:14.799]
Caesars order to get the views

[00:18:16.900]
I want where's very

[00:18:19.000]
small companies with less than a hundred employees

[00:18:21.200]
good 40%

[00:18:23.299]
disagree with that statement

[00:18:28.200]
we did a lot

[00:18:30.299]
more analysis and we're writing report

[00:18:32.400]
up on the results

[00:18:35.099]
but in general what

[00:18:37.700]
we're seeing is that according to this

[00:18:39.799]
data from this recent

[00:18:42.099]
assessment most day like today Ron and

[00:18:44.200]
Duke on premises were

[00:18:48.099]
staying at the dead legs are not there swamps

[00:18:50.099]
According to some gurus

[00:18:52.500]
at me in the Stream that

[00:18:54.700]
companies are able to maintain

[00:18:57.299]
high quality data in the state

[00:18:59.299]
of lakes and most importantly they're

[00:19:01.500]
not just for data scientists there

[00:19:03.700]
are graphical bi tools being used

[00:19:05.700]
heavily that provide fast for

[00:19:07.799]
a performance for Larry's an

[00:19:09.799]
exploration and finally that

[00:19:11.900]
their quality of data in The Lakes is suitable

[00:19:14.200]
for regular business users so

[00:19:17.000]
I must admit these results

[00:19:19.000]
in summary and it when we do have more

[00:19:21.099]
details in

[00:19:23.299]
the data were little bit surprising

[00:19:25.500]
to me but I think it's a good Testament

[00:19:27.799]
to how far and how

[00:19:29.799]
fast we become with

[00:19:31.799]
this new technology I do

[00:19:33.799]
but now the clouds

[00:19:35.400]
and I think that

[00:19:37.700]
is probably a good Segway to our

[00:19:39.700]
next speaker and Steve

[00:19:41.700]
will let you can talk about how

[00:19:44.099]
they're supporting both

[00:19:46.900]
regular users and power users in

[00:19:49.900]
Gator Lakes using their visual

[00:19:52.400]
bi to him a

[00:19:56.000]
couple

[00:19:59.500]
questions Wayne I'm curious

[00:20:01.599]
to know have you found or what's your take

[00:20:03.700]
on the people who were involved in

[00:20:06.000]
these projects in other words do you find

[00:20:08.200]
that the people who were in the data warehousing

[00:20:10.500]
team are the same people who are working

[00:20:12.500]
on data Lakes are they different teams

[00:20:14.799]
can you offer any contacts on that from your experience

[00:20:19.599]
yeah you know I think in

[00:20:21.700]
the early days a lot of the data

[00:20:24.000]
links were started by

[00:20:26.000]
Advanced analytics changed his

[00:20:29.500]
experiments to create

[00:20:31.500]
an analytical sandbox to fast-track

[00:20:34.000]
delivery station delivery

[00:20:36.400]
of analytical models predictive models

[00:20:38.500]
but will call you machine learning models

[00:20:40.900]
today I think very

[00:20:43.000]
quickly it's those things scaled-up or

[00:20:45.400]
failed a lot of them did not work

[00:20:47.599]
out but I

[00:20:49.700]
T took over that

[00:20:51.799]
infrastructure which makes sense

[00:20:53.799]
as an Enterprise environment

[00:20:55.799]
that can support

[00:20:57.900]
a lot of either a lot of

[00:20:59.900]
users the Enterprise or a very

[00:21:02.099]
important segments of users the power

[00:21:04.099]
users and data scientist

[00:21:06.200]
so administering that varmint

[00:21:08.299]
became Charlie

[00:21:12.900]
the domain of it

[00:21:14.900]
Steve's May

[00:21:17.000]
disagree with me but that's that's that's

[00:21:19.200]
what I've seen today

[00:21:22.099]
traveling teams

[00:21:30.700]
of these as he's Daylights have matured

[00:21:32.700]
take them over in the other group

[00:21:34.799]
so just mention like data governance 2

[00:21:36.799]
hours and you see the bi competency Center of

[00:21:38.900]
being involved in their search using standards

[00:21:41.200]
for these these platforms

[00:21:43.500]
as well so we'll talk more about that but

[00:21:45.500]
definitely is it's becoming mainstream woven

[00:21:48.400]
into the average few of

[00:21:50.500]
the organization that

[00:21:53.400]
a lot of organizations struggle

[00:21:56.099]
to reconcile their

[00:21:58.400]
expenditures on data warehouses and their

[00:22:00.900]
expenditures on Hadoop Hadoop

[00:22:02.900]
obviously this is less expensive by

[00:22:06.200]
terabyte and a lot of

[00:22:08.200]
business people look at the budget

[00:22:10.299]
on the bottom line of these environments

[00:22:13.099]
and you want to replace Adidas

[00:22:15.500]
warehouse but technically

[00:22:17.500]
that has not really been feasible

[00:22:19.500]
there are things that companies

[00:22:21.799]
are offloading from the data warehouse

[00:22:23.900]
that probably never belong there in the first place or

[00:22:27.700]
offloading me TL or detailed

[00:22:29.700]
data we're starting

[00:22:31.700]
to see this bifurcation at

[00:22:33.700]
least for now things to change

[00:22:36.099]
quickly that they don't houses is

[00:22:38.599]
well suited for supporting large

[00:22:40.599]
numbers of concurrent users we

[00:22:43.299]
need to do basic reporting and dashboarding

[00:22:45.900]
where is the date of lake is

[00:22:48.000]
suitable for power

[00:22:50.700]
users and for bi SWAT environments

[00:22:53.000]
to build things really quickly prototypes

[00:22:55.200]
an experiment test them deploy

[00:22:57.599]
them but now

[00:22:59.700]
we're starting to see a lot

[00:23:02.400]
standard applications

[00:23:04.599]
in ripped applications also happening in

[00:23:06.599]
Hadoop as well so

[00:23:08.799]
I think these

[00:23:11.000]
two environments or co-opting each

[00:23:13.000]
other there are quickly developing the

[00:23:15.299]
capabilities that the other one has and they're becoming

[00:23:17.799]
more and more identical will

[00:23:19.799]
never be the same about the

[00:23:21.900]
dividing line between them is it's getting

[00:23:24.299]
fuzzier but we are seeing

[00:23:26.400]
I do a birthday like taking over

[00:23:28.500]
and more and more of the functionality of Analytics

[00:23:32.200]
then I guess since people just kind of throw

[00:23:34.299]
this over to you real quick before you

[00:23:36.299]
jump into your presentation it really

[00:23:38.500]
you do want these two environments to

[00:23:40.500]
be coordinating collaborating

[00:23:42.500]
you want to be a lot of overlap between

[00:23:44.700]
them and it seems to me and

[00:23:46.799]
I know that you guys are kind of playing in this space

[00:23:49.200]
but from my perspective

[00:23:51.400]
and random in the endless space I'm

[00:23:53.599]
on the outside of all of this but

[00:23:55.799]
I fear was surgeons in business

[00:23:57.900]
intelligence almost like we

[00:24:00.200]
went down the road a big data analytics we

[00:24:02.400]
learned some interesting things but maybe

[00:24:04.500]
we're not as Tethered to the Core

[00:24:06.599]
Business objectives as world of business

[00:24:08.599]
intelligence was and I kind

[00:24:10.599]
of now see a Resurgence of use and

[00:24:12.700]
bi Tools in able to buy more

[00:24:15.200]
powerful infrastructure underneath

[00:24:17.200]
that can tap into traditional data

[00:24:19.299]
warehouse environment closest polling sites

[00:24:21.599]
from Dean Lakes in from these new

[00:24:23.599]
environments these is that what

[00:24:25.599]
you're saying or what your take on all that

[00:24:28.700]
yeah that's why I think the

[00:24:31.700]
power user I think there's terms

[00:24:34.099]
like citizen data scientist floating around

[00:24:36.099]
with your kind of interesting because a

[00:24:38.599]
lot of excitement around to do was

[00:24:40.599]
getting after all the granular

[00:24:42.700]
data there's not some it Department that's

[00:24:44.700]
pre-processing and

[00:24:46.700]
telling you what you should be analyzing it

[00:24:49.599]
it's sort of an expiration area but

[00:24:51.900]
I think you know what's been missing is

[00:24:54.000]
how do we give that same power to the

[00:24:56.000]
to the business users and then you

[00:24:58.099]
know you've got things like machine learning that are being

[00:25:00.099]
adopted by bi

[00:25:02.400]
and Alex schools out there that can speed

[00:25:04.599]
up the discovery process can put more power in the

[00:25:06.599]
hands of these power users are system data

[00:25:08.599]
scientist and things like that so I think it's it's

[00:25:10.599]
just been this Natural Evolution as head

[00:25:12.799]
of the next generation of native people

[00:25:17.099]
start using Technologies like to do

[00:25:19.200]
been filed there is I think

[00:25:21.299]
a generational

[00:25:27.200]
growth of different technologies

[00:25:29.500]
that need to keep up with the demands of these different types

[00:25:31.599]
of user so but I do see it coming back

[00:25:33.700]
to the end of the day sequel is

[00:25:35.799]
the language that people want to speak and

[00:25:38.200]
if you've got a gooey bass tool that can generate

[00:25:41.200]
sequel that can be utilized

[00:25:43.299]
by the date of our house or the date of Lake

[00:25:45.299]
I think that becomes a standard through

[00:25:47.500]
which you can do your houses so I

[00:25:53.400]
know here that assessment

[00:25:55.400]
which Wayne talked about from

[00:25:57.400]
which we got all that day that we were just sharing a

[00:25:59.500]
moment ago you will get a link

[00:26:01.700]
to that assessment in your follow-up email

[00:26:04.000]
later on this week so we hope you take

[00:26:06.099]
a look at that and Dive Right In and use

[00:26:08.299]
it really also not just understand

[00:26:10.500]
where you are going to see where you compare to other

[00:26:12.599]
companies and you can even do analysis

[00:26:14.700]
of a company's your size your region

[00:26:16.900]
and want some pork industry so it was

[00:26:18.900]
designed to provide some

[00:26:21.299]
really nice granular

[00:26:23.500]
detail to to give

[00:26:25.500]
you some perspective on where you are in your

[00:26:27.500]
organization and give you some advice on

[00:26:29.500]
which direction you should take so I think

[00:26:31.500]
it's a very powerful tool and I would recommend checking

[00:26:33.900]
out the intestines so with that Steve

[00:26:36.000]
would take it away

[00:26:38.000]
yeah thanks Eric I'm waiting and I'm really

[00:26:40.700]
pleased with the survey because I honestly

[00:26:42.900]
I've been in this industry 17

[00:26:45.099]
years overall I've been looking at the Juke

[00:26:47.400]
big data and data Lakes for like

[00:26:49.599]
the past I don't know 8

[00:26:51.700]
to 10 years I've never seen research frankly

[00:26:53.700]
that really gets into the adoption

[00:26:56.099]
the usage what platforms that are on day

[00:26:58.099]
lights that's really cool to see this

[00:27:01.000]
research going to come in the house and as

[00:27:03.099]
Eric said I think there's a lot more people out there

[00:27:05.099]
using its love to get the perspective from folks

[00:27:07.299]
but what I'd like to talk a little bit about is

[00:27:09.599]
what I've seen change in the technology I've

[00:27:13.000]
been at traditional

[00:27:15.200]
bi companies in my past I've

[00:27:17.400]
worked for large database companies

[00:27:19.700]
like teradata I've worked

[00:27:21.700]
at the Duke distribution vendors and

[00:27:23.900]
now at Arcadia day that we were really built to

[00:27:26.400]
focus on that challenge of how do we put

[00:27:28.400]
Power of bi into

[00:27:30.599]
the hands of people that want to go after these modern-day

[00:27:33.599]
to platforms if you will

[00:27:35.400]
and really what

[00:27:38.799]
we're starting to see now is that large Enterprises

[00:27:41.000]
as I mentioned he's bi competency

[00:27:43.200]
centers they are choosing new bi

[00:27:45.299]
standards for their data Lake which are separate

[00:27:47.599]
from and really not competitive

[00:27:50.000]
with their day to wear Warehouse

[00:27:52.099]
bi infrastructure because his Eric mentioned

[00:27:54.099]
at the beginning the

[00:27:56.200]
technology for bi that came out around you

[00:27:58.900]
know that was really based

[00:28:01.299]
on the processing power number in

[00:28:03.299]
things that we had been and I think there's a whole new world

[00:28:05.299]
around big data which is obviously

[00:28:08.099]
the size of it but also the variety

[00:28:10.099]
the date of the speed at which comes in the

[00:28:12.200]
need for to go have more real-time access

[00:28:14.200]
as well as just distributed

[00:28:16.200]
systems and

[00:28:19.099]
yeah the whole concept of figure

[00:28:21.700]
out what questions you need to ask and true date of

[00:28:23.799]
discovery versus again having the IT department

[00:28:25.799]
try to cheer a and

[00:28:27.799]
build cubes and things like that that are

[00:28:30.099]
based on business requirements but maybe not

[00:28:33.000]
opening up all the granny with detailed

[00:28:35.299]
data to the exploration of

[00:28:37.500]
some of these citizen data scientist

[00:28:39.700]
or power users want to do on

[00:28:41.799]
all these new sources of information they now have access to

[00:28:44.099]
so one with its way

[00:28:46.099]
to say that I think times are changing and there

[00:28:48.099]
is this inflection point and if you look at the technology

[00:28:50.400]
history is kind of interesting because

[00:28:53.000]
the date of Warehouse relational technology

[00:28:55.299]
is Eric mentioned was built at

[00:28:57.299]
the time when processing your Hardware

[00:28:59.400]
was really expensive memory is really expensive

[00:29:01.500]
and there's a lot of optimization done

[00:29:03.599]
at the stalker level to immigrate very very

[00:29:05.799]
tightly with the hardware to make sure you're maximizing

[00:29:08.500]
resource utilization so

[00:29:10.900]
those systems been to be proprietary which

[00:29:12.900]
is not a bad thing that I actually super high performance

[00:29:15.200]
but you couldn't take the

[00:29:17.700]
eye server software or a bi server

[00:29:19.799]
and run it in that same software

[00:29:22.500]
layer with a database it's running because it

[00:29:24.500]
was so you know

[00:29:26.500]
engineered for performance so

[00:29:29.000]
that's why you've got

[00:29:31.000]
traditional Vehicles spit on servers

[00:29:33.700]
desktops and will access

[00:29:35.900]
date in the date or out and there's nothing wrong with that that's

[00:29:38.000]
how it was set up so when you look at the analytical

[00:29:40.000]
process you've got to create

[00:29:42.700]
physical optimizations

[00:29:44.700]
of the date and how it stored physically on

[00:29:47.299]
disc and there's aggregate that

[00:29:49.299]
are created of course happy Samantha

[00:29:51.599]
Claire's at the bi to level

[00:29:53.599]
which connect to different data sources but

[00:29:55.799]
a lot of times that Dad

[00:29:57.900]
has to be secured and load it in two different places

[00:30:00.000]
and when you start talking about

[00:30:02.099]
realtime in but it's the laws

[00:30:04.099]
of physics date that you just going to be late and see

[00:30:06.099]
is your movie theater across the wire from one system to

[00:30:08.099]
the other not to mention the overhead of multiple

[00:30:11.000]
security layers and models and roll bass access

[00:30:13.099]
controls and need to be connected and

[00:30:15.299]
acceptance sync between these different system so when

[00:30:17.299]
you start to throw a big data into the architecture

[00:30:19.700]
you got semi-structured data you've

[00:30:21.900]
got these massively parallel systems

[00:30:24.000]
like to do and plowed object stores

[00:30:26.299]
and you

[00:30:28.599]
just the volume of date and the time it takes to move it

[00:30:30.700]
at that are you lose that ability to connect

[00:30:32.799]
natively and do real times and

[00:30:34.799]
now it's on the system so

[00:30:36.799]
when we found it Arcadia data

[00:30:39.000]
back in 2012 it was really to solve

[00:30:41.299]
that problem in for people that have the day like

[00:30:43.400]
in place how can we give large

[00:30:46.299]
numbers of concurrent business

[00:30:48.400]
users access that information and

[00:30:50.500]
the big I was

[00:30:53.000]
you rather than having to work

[00:30:55.400]
as a server let's do what

[00:30:57.500]
they said to do was all about it ring the

[00:30:59.500]
processing to the data let's build a bi

[00:31:01.599]
server that is fully distributed

[00:31:03.799]
runs and parallel across all the

[00:31:05.799]
Dayton OH so rather have separate bi

[00:31:07.799]
server we said let's use the servers

[00:31:10.000]
that are already in place will install and run

[00:31:12.099]
or software natively on each

[00:31:14.099]
of those mailers and we talked about needed bi that's

[00:31:16.299]
what we're talking about it I'm bi server that

[00:31:18.299]
takes advantage of the open

[00:31:20.299]
nature of Open Source software and

[00:31:22.400]
just modern-day to architectures like

[00:31:24.500]
the father he's got lots of processing engine

[00:31:26.599]
that can run on those data notes

[00:31:28.599]
and take advantage of the low-cost commodity

[00:31:30.599]
hard and you're

[00:31:32.900]
the resource utilization may not be as

[00:31:34.900]
Kylie optimize but the cops so

[00:31:36.900]
much lower you can just continue to throw

[00:31:38.900]
machines out of there fairly low cost

[00:31:40.900]
and scale extremely well that

[00:31:43.000]
was a big change that we made it on the architecture

[00:31:45.299]
side which also has

[00:31:47.299]
huge advantages from the overhead side where

[00:31:49.400]
you don't have to optimize physical layer

[00:31:51.400]
of spice you create a semantic there once

[00:31:53.700]
you can connect natively semi-structured data security

[00:31:56.900]
is done once we inherit security from

[00:31:59.099]
the underlined file system

[00:32:01.099]
and security systems

[00:32:03.099]
like a patchy century and Ranger and some of those and

[00:32:06.000]
you want to put in leggings

[00:32:08.000]
there you don't have to bring it into separate analytical

[00:32:10.299]
where which just by Nature gives you more real-time

[00:32:12.799]
access to the data

[00:32:14.299]
that's really architecture in the results

[00:32:16.799]
look like this this is a proof-of-concept

[00:32:19.599]
that we did from someone who's not a current

[00:32:21.799]
customer teleconferencing

[00:32:25.299]
platform but not allowed

[00:32:27.299]
to name but they're requirement was that

[00:32:29.799]
they needed 30 some current

[00:32:31.900]
customer success

[00:32:34.099]
manager is to be able to analyze the

[00:32:36.299]
log information around

[00:32:38.799]
the use of the spell conferencing service

[00:32:41.099]
to look for bottlenecks or issues

[00:32:43.900]
when service with would go bad and things like

[00:32:45.900]
that so they had to be complex queries it

[00:32:48.200]
had to be bi school and it had to

[00:32:50.200]
be 8:30 concurrent users and I

[00:32:52.900]
took away some of the names

[00:32:55.400]
of the different schools cuz I'm not trying to point out any

[00:32:58.500]
issues with seat want to get pensions but

[00:33:00.599]
to connect the the issue was there trying to

[00:33:02.599]
take a traditional bi2 on connect it to

[00:33:04.700]
a sequel on Hadoop engine

[00:33:06.700]
and there's three different engines they tried

[00:33:09.200]
in blue spray

[00:33:11.200]
and yellow and once you got above 5

[00:33:13.400]
concurrent users the performance

[00:33:15.500]
degraded significantly and results

[00:33:17.799]
were not returning so

[00:33:20.900]
again the contractors that Arcadia

[00:33:23.299]
data or I need to be ice platform is not

[00:33:25.599]
is she playing the gungeon it's

[00:33:27.599]
not just doing scans it's actually

[00:33:30.200]
optimizing performance and thinking like

[00:33:32.200]
a bi server that runs in the date

[00:33:34.299]
of platform that gives you that ability to support

[00:33:36.299]
lots of concurrent business users and

[00:33:39.099]
accelerate existing bi tools or

[00:33:41.200]
we provide our own v i II or II

[00:33:43.299]
and of course only

[00:33:45.599]
sit in the day too late cuz we talked about the data warehouse

[00:33:47.700]
is not going away it serves a very strong

[00:33:49.799]
purpose and we're close that didn't belong there are

[00:33:51.799]
moving on to other systems you've also got things

[00:33:53.900]
like event streaming was a really

[00:33:56.000]
popular now people want to be to stream data

[00:33:58.000]
from iot sensors out on the field and

[00:34:00.099]
connect the car so tell show demo

[00:34:02.500]
on the 2nd but needing to

[00:34:04.700]
be alerted to but also see

[00:34:07.299]
data as it's happening in real time respond

[00:34:10.199]
business that is happening with the ability

[00:34:12.300]
to drill to detail in the

[00:34:14.300]
data link or connect to other systems

[00:34:16.400]
where there's no sequel or

[00:34:18.699]
a relational system if you had a visual

[00:34:21.199]
eyes all that in one place is the

[00:34:23.400]
requirement of course if you would have any

[00:34:25.699]
bi to all I needed bi tools are

[00:34:27.900]
no different than can support that

[00:34:30.599]
so just the one last thing on Arcadia

[00:34:32.800]
specifically is the other thing we really

[00:34:34.800]
thought about was not

[00:34:37.300]
talked about cubes and all that you've seen

[00:34:39.300]
this idea that i t

[00:34:41.500]
Bill's bees with the business based

[00:34:43.599]
on business requirements in advance and they can be

[00:34:45.599]
fairly complex projects

[00:34:47.599]
to take on you build a cube and you're trying to teach

[00:34:49.599]
people how to fish but

[00:34:51.599]
you're only handing them a certain number of fishes within

[00:34:54.400]
that Cube and every time they ask for more

[00:34:56.400]
information you got to go in and then more fish or

[00:34:58.599]
recreate that Cube so what we talked about his

[00:35:00.900]
I can we give and users dry

[00:35:03.199]
nail or access to all the data for a Dock

[00:35:05.199]
Brewery in the data Lake and

[00:35:07.400]
provide optimizations

[00:35:09.500]
as we go on the fly so we treat

[00:35:11.800]
these things using some machine

[00:35:13.900]
learning and a recommendation engine Roxy looking

[00:35:15.900]
at one of the cruise it's people are right

[00:35:18.300]
what are the tables are accessing of

[00:35:20.300]
the files are accessing it will recommend

[00:35:22.699]
to the administrator what we call analytical

[00:35:24.900]
fuse and these are passing

[00:35:27.199]
mechanisms our guest and physical models

[00:35:29.599]
that will build back on disc in

[00:35:32.000]
the distributed file system run the product

[00:35:34.000]
store got to take advantage of memory

[00:35:36.300]
on the machines to make sure that

[00:35:38.400]
the next time those queers come in there's a crawl

[00:35:40.400]
space optimization decision which

[00:35:42.599]
will write that query to the fastest way to

[00:35:44.599]
bring it back so there's no modeling in advance

[00:35:46.900]
human is still involved in

[00:35:49.099]
cashews the physical modeling strategies

[00:35:52.000]
but it's it's using AI if you

[00:35:54.000]
will or machine learning said recommend the

[00:35:56.900]
best ways to speed

[00:35:58.900]
up those queries in the future so that's smart

[00:36:01.500]
acceleration is what we

[00:36:03.500]
call it when our system that again.

[00:36:06.400]
Cubes on her head nothing wrong with a lot Cubes

[00:36:08.500]
but it doesn't reduce don't want to wait

[00:36:10.599]
and see in the process

[00:36:12.800]
speaking of the process I'm going

[00:36:15.000]
pretty fast but you'll get besides afterwards

[00:36:17.400]
I think what we're seeing is if you look

[00:36:19.400]
at the bottom and white

[00:36:21.400]
A lot of people are taking it away

[00:36:23.400]
and they're treating it just like

[00:36:25.400]
another day to wear house or storage machine

[00:36:28.199]
and they're trying to take their bi

[00:36:30.400]
server connect to it there's nothing wrong

[00:36:32.400]
with starting that way but we see is this

[00:36:34.400]
analytical process that can really

[00:36:36.500]
be delayed because as

[00:36:38.599]
Wayne refer to if you're following

[00:36:40.800]
the end of model and third normal form and

[00:36:42.900]
at staging area yeah there's a process

[00:36:45.300]
by which you're going to land the data in The Lakes

[00:36:47.699]
you're going to transform it in some model trait

[00:36:50.000]
the scheme on that then you connect your bi

[00:36:52.099]
tool it's running on separate server just

[00:36:55.000]
the modeling part of that can take weeks so

[00:36:57.099]
this doesn't read before you can even

[00:36:59.099]
start to connect to be a server to

[00:37:01.199]
it you got this modeling us done and

[00:37:03.300]
then oh by the way you're going to trade cubes on

[00:37:05.400]
BI server to speed up performance

[00:37:07.400]
there when she brought in the date of from

[00:37:09.599]
the lake in step five and then

[00:37:11.699]
again you've got to secure it in two places go

[00:37:14.000]
before you get two steps fix it could be weeks

[00:37:16.099]
or months before you're actually able to do

[00:37:18.099]
any kind of analysis on day today

[00:37:20.199]
started out the date of eggplant before

[00:37:22.400]
you put in a production and maybe some additional modeling

[00:37:24.699]
so with the native approach again

[00:37:27.500]
one system where it stored and analytical

[00:37:29.800]
processing is done in there so

[00:37:31.800]
we land and secure at once you

[00:37:34.300]
can normalize and create schema if

[00:37:36.400]
she wants there is a semantic where they can also

[00:37:38.400]
connect to semi-structured data structure

[00:37:41.199]
raised in those types of things in your anal to the

[00:37:43.199]
Discovery process is much faster because

[00:37:45.699]
you're not moving data you don't need

[00:37:47.800]
to worry about the optimizations in advance

[00:37:50.099]
it'll run just fine for the

[00:37:52.099]
discovery queries and then that a

[00:37:54.099]
i driven from smiling the smart acceleration

[00:37:56.599]
it's done I can be after the fact

[00:37:58.800]
when you decide you want to put something

[00:38:00.800]
in the production so you're not moving the day that's one

[00:38:02.900]
security model and you're taking advantage

[00:38:05.000]
of next-generation technology to

[00:38:07.000]
speed up at analytical process as well as

[00:38:09.000]
Molly on the back end so greatly

[00:38:11.000]
accelerates that time the Insight

[00:38:13.000]
from weeks or months 2 days and I can tell

[00:38:15.000]
you again having work for big

[00:38:17.099]
database companies I've had customers I've worked

[00:38:19.199]
with you said your anytime we need to

[00:38:21.400]
add a new dimension to the schema

[00:38:23.599]
in the data warehouse is literally 6

[00:38:26.199]
to 12 months of time and a million dollars

[00:38:28.300]
of cost so if you just want to bring in I don't

[00:38:30.300]
know clickstream data into the warehouse for

[00:38:32.500]
Discovery lot of these systems and departments

[00:38:34.900]
have been fed up by which because it's

[00:38:36.900]
so highly governed it's just a long process before

[00:38:38.900]
you can get into some of that so I

[00:38:42.099]
think that's why you saw this need for data

[00:38:44.199]
scientists and the Resurgence of creation

[00:38:48.199]
of data Lakes words more exploratory nature

[00:38:51.800]
so I know we're going to save some time for questions I'm

[00:38:54.099]
going to be a quick demo flyby of kind of what's

[00:38:56.300]
possible with a date of native

[00:38:58.400]
technology come

[00:39:00.400]
back to this but

[00:39:02.800]
I slept over here and

[00:39:05.000]
sorry my emails up here but

[00:39:07.400]
this is Arcadia

[00:39:09.400]
data and in this instance here

[00:39:11.500]
we've created a couple of different Dental environments

[00:39:13.900]
I've got one on connect guitar this is cyber

[00:39:15.900]
security application that

[00:39:17.900]
I won't show here but you

[00:39:20.000]
go ahead and launch this and this is a a

[00:39:22.000]
demo environment talking about connected Vehicles

[00:39:24.400]
which it's a very hot topic now on

[00:39:26.900]
automated autonomous Automobiles

[00:39:29.400]
and you can imagine as a fleet

[00:39:31.500]
manager for let's say I don't know some

[00:39:33.900]
Service Company like AT&T is putting

[00:39:35.900]
Vehicles out in the world

[00:39:37.900]
do you want to get some notifications of things

[00:39:40.000]
that are happening so you can have real-time event streams

[00:39:42.300]
that are coming in this could be coming from something

[00:39:44.699]
like Apache Kafka could be coming in

[00:39:46.699]
from spark streaming people

[00:39:48.699]
you stole her and indices and

[00:39:50.800]
things like that for more real time updates

[00:39:53.000]
and analytics and we're requesting this information from

[00:39:55.500]
the vehicles and again

[00:39:57.500]
we're looking at your legal and

[00:39:59.500]
departures and yellow things in orange

[00:40:01.800]
or collisions in hazardous conditions are

[00:40:03.800]
in red and we're looking at you

[00:40:05.800]
a map that you can zoom in to and liquid

[00:40:08.699]
in San Francisco for specific events that

[00:40:10.699]
are happening or I can slick on an

[00:40:12.699]
individual then or

[00:40:14.800]
a car and doing more detailed

[00:40:16.900]
analysis so the history of

[00:40:19.000]
that car and what's been happening over time

[00:40:21.199]
so this could be across different drivers

[00:40:23.599]
and for this then we see all these different events

[00:40:25.699]
that happened we've got some scores

[00:40:28.099]
that are being calculated these are results

[00:40:30.199]
of Sparks jobs that are looking at

[00:40:32.500]
the acceleration of

[00:40:34.599]
Russian score if you over some for

[00:40:36.800]
this vehicle how much is it been accelerating what's

[00:40:39.199]
the strength that which these people have

[00:40:41.199]
been breaking steering

[00:40:43.199]
knobs are sensors

[00:40:45.800]
that are on the device for accelerometers and things

[00:40:47.900]
like that and then you can start to

[00:40:49.900]
do some correlation analysis for those drivers

[00:40:51.900]
are car and things like that and look at things

[00:40:53.900]
like it was there a correlation between

[00:40:55.900]
people that drive really aggressively in

[00:40:57.900]
the number of collisions at the raining again

[00:41:00.099]
obviously

[00:41:02.699]
but also gets into things like predictive

[00:41:05.000]
maintenance and what's the correlation between

[00:41:07.099]
acceleration and the

[00:41:09.400]
needing to replace brakes

[00:41:11.699]
or transmissions and

[00:41:13.699]
things like that so the fleet manager you've got

[00:41:15.699]
the ability to monitor things in real time but also

[00:41:17.699]
drill to detail looks for correlation and

[00:41:19.699]
all that within the simple UI

[00:41:23.099]
that's a quick flyby the types of things that

[00:41:25.099]
are possible I got

[00:41:27.599]
one question that came in about what industries

[00:41:29.699]
are these in data likes that is and

[00:41:31.699]
I think it we see it hugely and financial

[00:41:34.000]
services telecommunications

[00:41:36.400]
government's retail CBG

[00:41:38.900]
all the traditional industries that have lots

[00:41:40.900]
of products locks the customers dial

[00:41:44.300]
TN Spencer devices will be a lot more growth

[00:41:46.900]
and things like that but

[00:41:49.500]
it really does span all

[00:41:51.800]
different kinds of Industries in different forms of use

[00:41:53.800]
cases so just real quick I wanted

[00:41:55.800]
to show the tool itself and how easy

[00:41:58.099]
this build stuff so this is environment we've got

[00:42:00.199]
running it's connected to

[00:42:02.199]
the date I should

[00:42:04.300]
say city in a day like from the distribution

[00:42:07.000]
and all I want to do is show you

[00:42:09.099]
how I build a sports I'm going to connect to

[00:42:11.199]
data source this is just TV data

[00:42:13.800]
on a few worship across different

[00:42:16.199]
channels from a TV network I

[00:42:18.400]
called it a Kirsten TV cuz I know Wayne's

[00:42:20.599]
going to get into TV radio again someday

[00:42:22.699]
just kidding but

[00:42:26.000]
I'm just going to take that day to said that's already been

[00:42:28.000]
connected to I won't bore you with how we do the

[00:42:30.099]
connections but it connects the lots of different stuff and

[00:42:32.699]
now I'm going to build a dash for it so I just click the button

[00:42:34.699]
that says create dashboard it

[00:42:36.699]
pulls in the day that's been connected to this

[00:42:39.099]
is looking at session ID user ID

[00:42:41.099]
etcetera I just want to simplify that

[00:42:43.199]
down someone to look at it here I'm

[00:42:45.599]
going to look at

[00:42:48.400]
overtime so bringing the date string

[00:42:52.500]
call record count for all channels are

[00:42:54.500]
programs that

[00:42:56.599]
really quickly so now I've got a nice simple date

[00:42:58.699]
string I'm looking at all the record count overtime

[00:43:00.800]
for that but you know

[00:43:02.800]
where visualization tool so let's try something

[00:43:05.000]
we've got something like 30 different visualization

[00:43:07.300]
pipes and I'm lazy I don't want

[00:43:09.300]
to try and testing myself so I'm just

[00:43:11.300]
going to click on the spot and cut Explorer visuals

[00:43:13.300]
what this does is it used

[00:43:15.900]
to sew machine learning and best practices that

[00:43:18.099]
are built within the product to recommend

[00:43:20.400]
different visualization types based on the dimensions and

[00:43:22.699]
measures that I selected to

[00:43:25.000]
hear some different things like bubble charts and

[00:43:27.099]
Scatter Plots and horizontal bar

[00:43:29.300]
charts there's a calendar heat map which

[00:43:31.800]
is kind of interesting so I'll grab that and

[00:43:34.800]
this again it's just all records of her time

[00:43:38.000]
and that you can see the hot spots of days

[00:43:41.099]
in the months that were really heavy in

[00:43:43.099]
terms of people watching TV so we'd like to try

[00:43:45.199]
and explore that sauce closed

[00:43:47.199]
so David's Stone

[00:43:54.400]
and I'm going to splice

[00:43:56.800]
a little bit differently I'm going to look at

[00:43:58.800]
channels and programs

[00:44:02.000]
and the measures will stay as record

[00:44:04.000]
count bunny or limit that to the

[00:44:07.000]
top 50 just to speed up

[00:44:09.000]
looking at here

[00:44:11.000]
and sit down so refresh

[00:44:13.000]
that and

[00:44:16.900]
it's turning but there has got all channels

[00:44:19.300]
different programs and record count again nice

[00:44:21.500]
type of the form but I'd like to visualize it's

[00:44:23.800]
let's see what the system recommends to me

[00:44:29.000]
and this is the real data if you can visualize

[00:44:31.000]
it's not just make

[00:44:33.099]
up thumbnails I can actually see the results

[00:44:35.099]
here I'll go ahead and click the horizontal

[00:44:37.099]
bar chart and

[00:44:40.099]
that looks good from channels

[00:44:42.199]
and it's ranked and sorted

[00:44:46.000]
I got the top 15 so

[00:44:48.099]
save close that and

[00:44:52.199]
the one thing I want to do is

[00:44:54.199]
add a couple filters real quick and

[00:44:56.199]
I will open up the questions again just

[00:44:58.400]
showing how you can connect to data Explorer just

[00:45:00.599]
like you would expect within a high school so

[00:45:02.699]
it's asking filters

[00:45:05.199]
when I add a filter for a channel

[00:45:08.599]
and filter for programs

[00:45:13.699]
say that one more time see you

[00:45:15.699]
at so

[00:45:19.099]
here we have it and I've got some filters

[00:45:21.300]
stuff and I can see what's happening over all time

[00:45:23.300]
for all channels let's pick a channel

[00:45:26.699]
655 in here dork

[00:45:34.300]
anorexia looks this once let's see what

[00:45:36.300]
are the top shot what are the top programs on Syfy

[00:45:38.599]
show face off

[00:45:40.699]
Friday Night SmackDown

[00:45:43.099]
Bourne Ultimatum X-Men

[00:45:45.699]
things

[00:45:55.400]
like that so, but

[00:45:57.400]
gives you a sense of how you can do this again

[00:45:59.599]
Arcadia data running directly in

[00:46:01.599]
the day delay giving you access call The Grind of a data

[00:46:03.699]
so that I will stop

[00:46:05.699]
jabbing and we do

[00:46:09.500]
we do have some good questions here so let me just

[00:46:11.500]
start throwing some over to you one

[00:46:13.800]
of these pennies is asking about data quality

[00:46:16.199]
where the quality updated being curated

[00:46:18.500]
inside the crkt architecture

[00:46:23.000]
we do not focus on that

[00:46:25.400]
a prep that's something that are Partners like

[00:46:27.500]
trifecta box. A strange

[00:46:29.599]
that some folks like that will get into

[00:46:31.800]
we have a little bit of day to prep stuff within

[00:46:33.900]
it for the the business analyst what

[00:46:36.000]
we were really glad those Partners

[00:46:38.000]
provide a solution

[00:46:40.199]
that runs within the date awake to do all

[00:46:42.199]
the standard preparation steps that you

[00:46:44.199]
would want for more curated data

[00:46:47.300]
okay and working

[00:46:49.300]
pens is asking about 3

[00:46:51.800]
as a possible destination can you kind

[00:46:53.900]
of talk about your relationship

[00:46:56.400]
with Amazon at 3

[00:46:59.000]
yeah absolutely we have a number of customers

[00:47:01.500]
that are fully on the cloud

[00:47:04.900]
trying to wish names I can mention I didn't do

[00:47:06.900]
stars one Turner Broadcasting

[00:47:08.900]
is another but a lot of people are

[00:47:10.900]
starting to store data directly in S3

[00:47:13.800]
to still leveraged you could do because in many

[00:47:15.800]
cases so I can be able to run in the elastic

[00:47:18.199]
tear and connect directly to data and S3

[00:47:20.400]
to visualize it but

[00:47:22.500]
that's that's something we've had for a while we

[00:47:24.699]
just announced support for Microsoft

[00:47:26.800]
is your data Lake store as

[00:47:28.800]
well okay you

[00:47:30.800]
must have been reading my mind cuz that was my next

[00:47:32.800]
question I was asking doesn't

[00:47:35.800]
work in Microsoft as you're in the answer is

[00:47:37.800]
now yes over

[00:47:41.900]
to you this is an interesting one we got to talk about

[00:47:44.000]
it already but one of the attendees nose than

[00:47:46.099]
likely it

[00:47:48.800]
people used to working on a date of Warehouse

[00:47:50.900]
are going to require better than mine

[00:47:53.300]
step shift what

[00:47:56.300]
did You observe that can facilitate them to

[00:47:58.500]
count every Orient themselves to

[00:48:00.800]
focus on supporting a data

[00:48:03.000]
link versus the date of our house

[00:48:07.699]
well I think

[00:48:09.800]
it looks good is that a lot of the skills that

[00:48:11.900]
those people in there they're dbas

[00:48:14.000]
or what have you have her to

[00:48:16.500]
please reusable I think we're starting to

[00:48:18.500]
see more and more analytical were close also

[00:48:20.500]
moving to the dead awake for new applications

[00:48:23.099]
as people want to build them in as

[00:48:25.300]
one of the colors ask about you know Theatre

[00:48:27.500]
quality cleansing skiing all those things are

[00:48:29.599]
still really valuable and important I think was

[00:48:32.300]
changes just rethinking what's

[00:48:35.500]
available in terms of bi tools I think you

[00:48:37.900]
know we were first to

[00:48:39.900]
Market to be to connect two things like apache-kafka

[00:48:42.000]
natively because we're just kind

[00:48:44.099]
of in that space and they've got a new case

[00:48:46.199]
equal in her face that allows you to to

[00:48:48.599]
query streams of information or

[00:48:50.699]
things like Apache solr or

[00:48:52.699]
Apache kudu and other types of data platforms

[00:48:55.199]
that have some benefits to it and being able

[00:48:57.199]
to explore data and take advantage of

[00:48:59.400]
nested data like things

[00:49:02.099]
like Jason's trucks and raised where you

[00:49:04.099]
got the meditate at in the the

[00:49:06.400]
date of hornets so you may not

[00:49:08.400]
need to build a lot of scheme

[00:49:10.500]
in advance just give users more access

[00:49:12.900]
to it but you still need to have things like

[00:49:15.000]
roll base access control and security and things

[00:49:17.099]
like that and I

[00:49:19.099]
think those concerns about securing

[00:49:21.199]
all those involved installed by the community

[00:49:23.400]
I think the next wave is just providing

[00:49:25.699]
tools that can take advantage of those who brought

[00:49:28.099]
her stuff so I don't know if that answers the

[00:49:30.199]
question I think Wayne probably gets more involved and

[00:49:32.500]
used appliances of training and education

[00:49:34.500]
on

[00:49:39.000]
yeah I mean I was going to say that

[00:49:41.000]
you still hear you got to pay the piper at

[00:49:43.000]
some point and you have to create

[00:49:45.000]
a schema for this data I

[00:49:47.099]
mean the value of the day too late for power

[00:49:49.199]
users it was scheme on reading

[00:49:51.199]
didn't have to wait for my Tita model it but

[00:49:54.599]
you know at

[00:49:56.599]
some point especially when you're trying to get

[00:49:58.599]
strong query performance for

[00:50:00.599]
large numbers of concurrent users you

[00:50:03.199]
probably do want to model the data and

[00:50:05.699]
that raises the question I had for Steve

[00:50:07.900]
when you talk about your smart

[00:50:10.199]
acceleration you kind of insinuated that

[00:50:12.199]
you really didn't need to model the

[00:50:14.199]
data that using machine learning your

[00:50:18.599]
tool would be

[00:50:21.199]
able to eventually

[00:50:23.500]
crate Autumn Cash's

[00:50:26.199]
I'm Aggregates automatically

[00:50:29.599]
so that you can get up and running

[00:50:31.800]
pretty quickly like in a matter

[00:50:33.900]
of days without having

[00:50:35.900]
to do any modeling at all the tool wood would

[00:50:38.199]
essentially create structures

[00:50:40.199]
on the Fly based

[00:50:42.400]
on queries you feed it may

[00:50:44.400]
be priming the pump to

[00:50:46.699]
deliver the decline of performance

[00:50:48.699]
that users would want the morning if that's accurate

[00:50:51.199]
reflection of what you're told us

[00:50:54.099]
yes it can certainly do it that way

[00:50:56.300]
you know but it's not pixie

[00:50:58.599]
dust right I mean you still need if you're

[00:51:00.699]
going to have your metadata definitions

[00:51:02.800]
and data Steward data catalogs business

[00:51:05.500]
terms in the date of

[00:51:07.500]
the tables that people want to access I think there's

[00:51:09.800]
you're still need that at some level

[00:51:11.800]
for tickly as once you done some initial exploration

[00:51:14.000]
if you want to provide a

[00:51:16.000]
broader View to a broader set of people having

[00:51:18.500]
those definitions of magic

[00:51:20.599]
players and things like that in place are also important so

[00:51:22.900]
anything that someone's built in the hive metastore we

[00:51:25.800]
going to do for other places can we take advantage

[00:51:27.800]
of that or if they've got to be at a cow in place

[00:51:29.900]
we conserve read from that and make it that

[00:51:32.000]
also available but yes I think even

[00:51:35.199]
for those query than that may have been to find

[00:51:37.300]
her the tables have been set now

[00:51:39.300]
there's going to be acceleration strategies

[00:51:41.800]
based on actual usage that

[00:51:44.300]
is in the ministry or may not think about an advance

[00:51:46.599]
so we can kind of monitor that in

[00:51:48.599]
the system will recommend other ways to

[00:51:50.800]
speed up the screws in the future so yes

[00:51:53.400]
it can be used turn on Raw data as it's

[00:51:55.500]
come in without any set up in advance but

[00:51:58.300]
it's also beneficial to have more

[00:52:00.400]
the curated date of that will live in support

[00:52:04.300]
soon as end user applications were you talking

[00:52:06.300]
about hundreds of thousands of users on

[00:52:08.800]
the system as well

[00:52:10.599]
but you don't necessarily require

[00:52:13.099]
it certainly wouldn't hurt for

[00:52:15.400]
users to create

[00:52:17.400]
a schema inside

[00:52:19.500]
I do using high for whatever

[00:52:21.599]
right to support Paris

[00:52:24.199]
but there's value to

[00:52:26.199]
it and we can also read it and take advantage

[00:52:28.199]
of it yes it's not required

[00:52:30.599]
yeah

[00:52:33.099]
I mean that seems to be the trend these days

[00:52:35.199]
with a lot of these new technologies and tools

[00:52:37.599]
is that processing

[00:52:40.599]
Powers is so great

[00:52:43.099]
that they can deal with now

[00:52:45.599]
the source chemo sloppy

[00:52:48.500]
scheme of it comes from the store soon

[00:52:50.599]
and do something with it and

[00:52:54.000]
give value pretty quickly and

[00:52:57.000]
you can only enhance that value by

[00:52:59.000]
by doing more to sign

[00:53:01.199]
up front and and

[00:53:03.199]
in your tool you actually help

[00:53:06.099]
do that as well with the smart accelerator

[00:53:08.599]
capabilities

[00:53:14.800]
okay good we got a couple more things here

[00:53:16.900]
are several accidentally throw and you kind of give

[00:53:19.000]
did the loot at this moment to go but there's a specific

[00:53:21.199]
question about state of catalogs

[00:53:23.699]
and semantic layers and what you were saying

[00:53:26.099]
is that Arcadia is startups

[00:53:28.699]
leverage though how

[00:53:31.800]
that happened in and where it happens

[00:53:34.099]
in the process of play

[00:53:39.599]
you cut out a little bit there I think you're asking

[00:53:41.599]
where did did catalogs

[00:53:44.300]
play with in all this or where does Archie

[00:53:46.300]
how

[00:53:51.800]
does that actually work

[00:53:56.900]
well then

[00:53:59.800]
pausing just cuz I wanted to so there's this consorcia

[00:54:02.500]
for what it's work at worth that were part

[00:54:04.500]
of the club make big day to work includes

[00:54:06.500]
vendors like tripac

[00:54:08.800]
the streams that's in water line data water line

[00:54:10.900]
is a bit a catalog that was built specifically

[00:54:12.900]
for data legs and it deep

[00:54:15.099]
in particular but there's also Elation date and

[00:54:17.099]
others that are out there so I'm

[00:54:19.099]
not an expert on those things but I understand

[00:54:21.099]
more more people is there

[00:54:23.300]
you know having multiple systems like the date

[00:54:25.400]
of Ross and the data link together they need,

[00:54:27.400]
definitions of a customer

[00:54:29.500]
and things like that and where that is stored in and

[00:54:31.599]
what day is available where is so we

[00:54:35.099]
can connect to any of those used car

[00:54:37.800]
buy back to the business user

[00:54:39.900]
the those definitions that have been to

[00:54:41.900]
find access that day to bring it in things

[00:54:44.300]
like that and then we got our own to all we have is semantic

[00:54:46.599]
layer which runs directly

[00:54:48.599]
to sort of in the tool in Hadoop and

[00:54:50.900]
business hours can create their

[00:54:52.900]
own data definitions for four

[00:54:55.099]
tables or data they're looking at that hasn't

[00:54:57.400]
been to find it and there could be no

[00:54:59.699]
user a in sales going to

[00:55:01.699]
name the date of one thing that makes sense

[00:55:03.699]
to them and use your beat it to spin

[00:55:05.900]
I don't know engineering might like name it

[00:55:07.900]
something else so that you can also do that at the bi

[00:55:10.000]
to level but there's obviously some

[00:55:13.400]
you know concerns

[00:55:15.599]
without if your data

[00:55:17.800]
governance purest and can I have in a single definition

[00:55:19.900]
for forgetting things like that

[00:55:21.900]
but yeah there's just going to all possibilities and

[00:55:24.000]
I would encourage people that go out and check out make

[00:55:26.000]
spaghetti work we've done a kind of a webinar

[00:55:28.000]
education Series in around

[00:55:30.199]
data catalogs and things like that and kind

[00:55:32.300]
of this world okay

[00:55:35.199]
good and here's a good question from

[00:55:37.199]
an attendee I think I know the answer but if

[00:55:39.800]
you would share with the audience from your perspective

[00:55:41.900]
what's the main difference or

[00:55:44.099]
differentiating feature between what

[00:55:46.099]
Arcadia doing and what time you could do with a

[00:55:48.099]
product like Tablo

[00:55:52.599]
yeah I need a different the key differentiating

[00:55:54.800]
feature is

[00:55:57.000]
the fact that were a massively parallel system

[00:55:59.500]
that runs directly with the data Tableau

[00:56:02.199]
can cluster environments

[00:56:04.300]
but our perspective

[00:56:07.099]
is Bennett there's a lot of knowledge about

[00:56:09.199]
how data stored on the on the individual

[00:56:11.599]
nose and are soft for sitting there next

[00:56:13.800]
to the data we can take advantage of

[00:56:16.099]
that that local knowledge or not just passing

[00:56:18.099]
sequel back and forth through

[00:56:20.300]
an odbc driver something like that

[00:56:22.300]
I'm working to run natively where it's

[00:56:24.300]
at so that just gives us tremendous

[00:56:26.400]
scale performance it's

[00:56:28.599]
a lower TCO solution overall

[00:56:33.099]
yeah that I didn't is the architecture that really

[00:56:35.199]
makes the difference but then also as

[00:56:37.400]
I talked about that process it really speeds up at

[00:56:39.400]
time the inside because you don't have any data latency

[00:56:41.699]
over-the-wire you're not needing

[00:56:44.300]
to move data from one system to another there's

[00:56:46.599]
the security where we

[00:56:48.599]
just inherited directly from the date of platform

[00:56:50.800]
you don't have to read Minister it in a separate

[00:56:52.800]
the guys who own that's just it's the

[00:56:55.300]
philosophy of a native the eye solution

[00:56:57.400]
which I think is becoming a thing

[00:57:02.699]
okay good and

[00:57:04.800]
you run both on private them

[00:57:06.900]
in the cloud right can you talk to that real quick yeah

[00:57:09.800]
absolutely I mean you

[00:57:12.699]
can get in a long debate on the difference difference

[00:57:15.099]
between spotting on Prem do us to

[00:57:18.900]
like customers I should stay just a deployment preference

[00:57:21.699]
a lot of people that go to the Pod

[00:57:23.699]
off and start out because they just don't want

[00:57:25.699]
I'm not in your man around data center so

[00:57:27.699]
you can star software just

[00:57:30.300]
as you would anything else in that environment there's

[00:57:33.000]
also some advantages that that

[00:57:35.099]
we have in it that obvious environments that I

[00:57:37.099]
won't get into on this but

[00:57:39.199]
I think there are some it

[00:57:41.300]
was virtual machine instances and things like that

[00:57:43.300]
someone some different thinking around

[00:57:45.800]
how you architect software to run

[00:57:47.900]
in those environments to scale precisely

[00:57:50.900]
with the work clothes I'm just going

[00:57:53.000]
to leave it at that but there's a lot of things

[00:57:55.099]
that we do in the positive very interesting I

[00:57:57.800]
think I break down and people

[00:57:59.800]
that are on 5:00 somewhere or

[00:58:01.800]
what in the survey

[00:58:04.099]
results thus far from Wayne up in

[00:58:06.199]
roughly 20% cloud

[00:58:08.599]
and a large majority still on Prime

[00:58:10.800]
but certainly a lot of people interested in the

[00:58:13.099]
hybrid and environments

[00:58:15.099]
but yes we can run there right

[00:58:27.000]
someone is asking is it similar

[00:58:29.000]
to what to know does your key

[00:58:31.300]
is giving direct access to the data

[00:58:33.300]
through this highly parallelize environment

[00:58:35.500]
right literally taking the processing

[00:58:37.699]
to the data and a highly parallel way

[00:58:39.699]
but you don't need to do virtualization

[00:58:42.099]
is that right

[00:58:44.099]
correct and you

[00:58:46.099]
know there's a need for data

[00:58:48.400]
virtualization or a value to it I

[00:58:50.400]
think for us you know where you want

[00:58:52.500]
now the physical copies in one place

[00:58:54.599]
like that's where you going to get the the huge

[00:58:56.599]
performance staying

[00:58:59.500]
so obviously we live in a world where

[00:59:01.699]
did it's all over the place or theirs needs for Federation

[00:59:04.000]
virtualization those types of things but I

[00:59:06.000]
think for production applications

[00:59:08.000]
where do you want to play at the hundreds of thousands of users

[00:59:10.099]
again that's why you would look at something

[00:59:12.199]
like a native architecture in addition

[00:59:14.300]
to the benefits of exploration

[00:59:16.500]
everything we just talked about

[00:59:19.500]
okay good thank

[00:59:24.599]
you Steve and thanks Wayne great stuff will

[00:59:26.800]
talk to you next time thank

[00:59:31.800]
you Erika and thank you Wayne what a great

[00:59:34.000]
presentation and thanks to our attendees for being so

[00:59:36.099]
engaged in everything we do and all the great questions

[00:59:38.099]
that have come in and just a reminder I will send

[00:59:40.099]
a follow-up email by end of day Friday

[00:59:42.400]
with links to the recording and

[00:59:44.400]
links to the assessment for you

[00:59:46.500]
and I will

[00:59:48.500]
see if I can get you a link to the additional demos

[00:59:50.500]
and such from arkadata so

[00:59:53.000]
thanks everybody and sponsoring

[00:59:55.900]
today's webinar I hope you all have a

[00:59:57.900]
great day