Developer Data Modeling Mistakes: From Postgres to NoSQL
Big Data
1. Big Data, Analytics and the
Future of Data Centres
Where are we and where are we going?
VMUG Meeting, Manchester February 12, 2013
Tony Lock – Programme Director
tony@freeformdynamics.com
www.freeformdynamics.com
www.freeformdynamics.com
Copyright 2013 Freeform Dynamics Ltd 1
2. About Freeform Dynamics
Industry analyst firm
Track IT industry developments and offerings
Track the evolution of IT related activity and needs in business
Advise both end user organisations and suppliers
Research approach
IT vendor and service provider briefings
Large scale studies - face to face, telephone and online
Community research programme
Investigate strategy, business case, architecture, best practice
Vendor patronage model allows free distribution
Media partnerships for both input and output
Copyright 2013 Freeform Dynamics Ltd 2
3. Agenda
Big data and Analytics
Where are we today?
The evolution of the data centre
Visions of the future
Will we ever reach Nirvana?
Closing thoughts
Copyright 2013 Freeform Dynamics Ltd 3
4. Big Data
What is it?
Is anyone doing it?
The only game in
town?
Copyright 2013 Freeform Dynamics Ltd 4
5. Defining Big Data?
Analogies
Panning for gold
Finding the needle in the hay stack
Identifying a weak signal in a very noisy environment
Find valuable patterns, trends, correlations, etc. in noisy,
unstructured, often complex, and high volume data sets
Doing analytics better / differently?
Copyright 2013 Freeform Dynamics Ltd 5
6. How much do you agree or disagree with the following
statements?
0% 20% 40% 60% 80% 100%
The emergence of advanced storage, access and
analytics solutions means the end of the
traditional RDBMS
Regardless of substance and reality of emerging
technologies and techniques, the term ‘big
data’ is currently being over-hyped by IT
vendors in an unhelpful way
I have a clear understanding of what the term
‘big data’ means
5-Totally agree 4 3 2 1-Totally disagree Unsure
Copyright 2013 Freeform Dynamics Ltd 6
7. The three Vs of Big Data
High physical volumes with low value
density
Rule of thumb
Volume
Generally think
of Big Data Different sources and formats or
when two of Variety information
these three
apply
Velocity Rapid rate of data movement, generation
or acquisition
Copyright 2013 Freeform Dynamics Ltd 7
8. The concept of value density
Low value density High value density
Traditional BI and data
Structured warehousing Structured
BIG
DATA
Traditional search and Unstructured
Unstructured document management
Low value density High value density
Copyright 2013 Freeform Dynamics Ltd 8
9. What level of growth are you seeing in the following
types of data within your organisation?
0% 20% 40% 60% 80% 100%
Structured data
(e.g. tabular data in RDBMSs)
Unstructured data
(e.g. documents, messages,
multimedia content, etc.)
5 (Extremely high growth) 4 3 2 1 (No growth)
Copyright 2013 Freeform Dynamics Ltd 11
10. In what form is your organisation’s most valuable or critical data
held (i.e. your crown jewels in information terms)?
0% 10% 20% 30% 40% 50%
Exclusively structured
Mostly structured
Equal split
Mostly unstructured
Exclusively unstructured
Copyright 2013 Freeform Dynamics Ltd 12
11. How is this changing?
0% 20% 40% 60% 80% 100%
Steady No Steady
shift change shift
(21%) (43%) (25%)
Rapid shift
Rapid shift
towards value in
towards value in
unstructured
structured data
data (7%)
(4%)
Copyright 2013 Freeform Dynamics Ltd 13
12. To what degree does your organisation exploit its
information assets for analysis and decision making?
0% 20% 40% 60% 80% 100%
Structured data
(e.g. tabular data in RDBMSs)
Unstructured data
(e.g. documents, messages,
multimedia content, etc)
5 (Fully) 4 3 2 1 (Very poorly)
Copyright 2013 Freeform Dynamics Ltd 14
13. Use of traditional and emerging technologies
Current level of use Change over next 3 years
0% 20% 40% 60% 80% 100% -60% -40% -20% 0% 20% 40% 60%
Legacy databases and file systems
General purpose RDBMS servers
High performance RDBMS configurations
OLAP multi-dimensional database systems
Write once read many (WORM) databases
Rule-based stream processing engines
In memory databases
Scale-out storage architectures
Distributed indexing and search
Distributed data analytics engines
5 (Extensive use) 4 3 2 1 (Not used at all) Unsure Less use More use
Copyright 2013 Freeform Dynamics Ltd 16
14. How much do you agree or disagree with the following
statements?
0% 20% 40% 60% 80% 100%
Developments in advanced storage, access and analytics are allowing
us to tackle problems today that were either too hard or too…
Developments in advanced storage, access and analytics are allowing
us to take different and better approaches to tackling some key…
Vendors and consulting firms are well geared up to providing us with
the support and services we need to take advanced storage, access…
We have a clear idea of the business benefits available to us through
the use of big data technologies and solutions
We have a clear idea of the advanced data storage and big data
analytic technologies that are becoming available
Database vendors are well geared up to support their customers with
appropriate licensing and commercial arrangements as data related…
5-Totally agree 4 3 2 1-Totally disagree Unsure
Copyright 2013 Freeform Dynamics Ltd 17
15. Elephants in the room
Access to data – what data is there and where is it?
Are there governance / regulatory / legal restrictions in play
concerning certain data sets?
Skill shortage?
IT skills
Numerical skills in user base
Just what questions could ‘Big Data’ help with?
How do we exploit any results we generate?
Feedback into ‘mainstream systems’
Copyright 2013 Freeform Dynamics Ltd 18
16. Data Centre Evolution
Where are we now?
‘Perfect Visions’
Will we ever get to Nirvana?
Copyright 2013 Freeform Dynamics Ltd 19
18. How much have you virtualised the following elements of your IT
landscape?
0% 20% 40% 60% 80% 100%
Your x86 server
estate
Storage infrastructure
Your corporate
network
Your desktop
environment
Totally Extensively Partially A bit Not at all Unsure
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 21
19. How much have you architected your IT infrastructure in the
form of shared resource pools such a private clouds?
Unsure Totally
6% 4%
Extensively
11%
Partially
20%
Not at all
40%
A bit
19%
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 22
20. YOUR VISION FOR DATA CENTRE COMPUTING
Following charts based on this question:
Putting all of the existing constraints and the current
state of the industry to one side for a minute, how
desirable would the following be as part of your perfect
IT vision?
Copyright 2013 Freeform Dynamics Ltd 23
21. THE CLOUD HOSTING THING (Perfect IT Vision)
0% 20% 40% 60% 80% 100%
All/most of our IT
requirements would be
fulfilled via hosted cloud
services
All/most of our own IT
infrastructure would be
based on private clouds
Highly desirable-5 4 3 2 Not at all desirable-1 Unsure or N/A
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 24
22. THE WHOLE CLOUD THING
0% 20% 40% 60% 80% 100%
All/most of your IT
requirements are fulfilled
via hosted cloud
services
All/most of your in-house
(or co-located) IT
infrastructure is based
on private cloud
architecture
Already there Within 1 yr Within 3 yrs Within 5 yrs
Within 10 yrs Later Never Don't know
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 25
23. OPERATIONS & MANAGEMENT (Perfect IT Vision)
0% 20% 40% 60% 80% 100%
We’d be able to provision workloads and new systems capacity on our
private clouds with a few clicks on a management console
Automation would mean that shared resource usage was continually
optimised as demands fluctuate, with no human intervention
We’d be able to migrate applications and workloads back and forth
between public and private clouds with ease
Hybrid cloud management capability would allow us to take an agnostic
view of resources, mixing and matching internal and external capacity
freely
We would have end-to-end visibility across on-premise and hosted
systems for management and troubleshooting purposes
We would have a consistent/joined-up way of managing security and
access across in-house and hosted systems
We would have a consistent/joined-up way of managing and protecting
data across in-house and hosted systems
We’d be taking a unified approach to operations and management across
servers, storage, networking, security, etc
Highly desirable-5 4 3 2 Not at all desirable-1 Unsure or N/A
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 26
24. OPERATIONS AND MANAGEMENT
0% 20% 40% 60% 80% 100%
New workloads and systems capacity are generally provisioned
with just a few clicks on a console
Use of key shared resources is continually optimised as
demands fluctuate, with no human intervention
You can quickly and easily move workloads back and forth
between public and private clouds
An agnostic view of resourcing means choices between internal
and external deployment are made purely on requirements and
fitness for purpose
You have end-to-end operational visibility across on-premise and
hosted systems
You have a consistent/joined-up way of managing security across
in-house and hosted systems
You have a consistent/joined-up way of managing data across in-
house and hosted systems
Already there Within 1 yr Within 3 yrs Within 5 yrs
Within 10 yrs Later Never Don't know
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 27
25. ORGANISATION & SERVICES (Perfect IT Vision)
0% 20% 40% 60% 80% 100%
Server, storage, networking, security and other specialists
would be working together seamlessly as part of a fully
integrated ops team
Other IT teams (developers, testers, support staff,
departmental IT, etc) would have self-service provisioning
capability to obtain IT resources
Non-technical users would have self-service provisioning
capability for new applications and services
IT activity and investment would revolve around the concept of
business services rather than systems
We would be able to easily and accurately charge or report IT
costs back to the business based on activity or consumption
Highly desirable-5 4 3 2 Not at all desirable-1 Unsure or N/A
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 28
26. ORGANISATION AND SERVICES
0% 20% 40% 60% 80% 100%
Self-service provisioning is in
place for IT teams
Self-service provisioning is in
place for end users
IT activity and investment
revolves around the concept
of business services rather
than systems
You can easily and accurately
charge or report IT costs back
to the business based on
consumption
Already there Within 1 yr Within 3 yrs Within 5 yrs
Within 10 yrs Later Never Don't know
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 29
27. How much are the following standing in the way of progressing
towards the vision?
0% 20% 40% 60% 80% 100%
Inability of suppliers to deliver on visions and promises
Lack of interest/appreciation from senior management
The business not ready to upset the status quo
IT not ready to upset the status quo
Cultural impediments to investment in shared infrastructure
Ingrained IT funding models don’t support new ways of
working
Historical under-investment means the mountain’s too high to
climb
Lack of a formally defined vision and strategy
We simply don’t have the time, resources or budget to focus
on anything other than short term priorities
Big impediment Significant challenge Minor challenge Not a problem Unsure
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 32
28. When it comes to driving forwards, how is it playing out, or
how do you think it’s going to play out in your organisation?
Single big
transformational initiative
to modernise things Build a modern
Other across the board environment for new stuff,
Things will be left largely 4% 4% and leave older systems
as they are for the
where they are
foreseeable future
11%
18%
Build a modern
Just let adoption of
environment for new stuff,
modern architectures,
and migrate older
tools and techniques
systems into it
creep along in an ad hoc
incrementally
manner
43%
20%
Enterprise 481 respondents
Copyright 2013 Freeform Dynamics Ltd 33
29. Summing Up
Can you have you cake
(OK Apple) and eat it?
Copyright 2013 Freeform Dynamics Ltd 34
30. Moving forwards
Senior business • Understanding is needed
• Communicate – often and in language business users can understand
awareness raising • Ensure governance processes are ‘big data aware’
Minimisation of ad • Proactive local involvement in planning/prioritisation
• Don’t forget about getting ‘big data’ derived results back into business use
hoc adoption • Clear policies/discipline around data usage
• Monitor usage patterns to spot trends early
Identify and deal with • Don’t get hung up on transient fads and fashions
real business issues • Embrace, substitute or block more persistent activity
• More of an orchestration approach to IT leadership
Facilitate flexibility via • Architect systems with hard core and flexible edges
core IT • Virtualise the edge to handle diversity and personal use
Copyright 2013 Freeform Dynamics Ltd 36
31. Adoption attitudes and tactics
Resist new ideas and solutions
Recipe for
Allow new stuff to creep in passively
disappointment
Ad hoc opportunistic adoption
Draw a line, target new apps only
Recommended
Establish beachhead, then expand by most early
movers
Big bang migration of everything
Copyright 2013 Freeform Dynamics Ltd 37
32. Available for download now:
Big Data and Analytics
Dazzling new solutions or irritating new hype?
Available from www.freeformdynamics.com
http://www.freeformdynamics.com/fullarticle.asp?aid=1590
A Vision for the Data Centre
Are you a Mover, Dreamer or Traditionalist?
Available from www.freeformdynamics.com
http://www.freeformdynamics.com/fullarticle.asp?aid=1604
Copyright 2013 Freeform Dynamics Ltd 38
MB to modify to tune to how she likes to define the BD problem
Emerging foundation for consistent pundit/vendor viewA bit contrived, but useful, and need to know because the 3Vs and becoming increasingly prominent
When you look at some of the information feeds, including structured (logs, M2M, etc) and unstructured (social, web), along with unstructured internal sources, you find a lot of them are very ‘dilute’ when it comes to valueThis these are the areas put into the ‘too difficult’ or ‘not cost effective’ to solve categoryThis is where parallel, distributed scale-out architectures – aka Big Data come inStress that it is not just about unstructured – e.g. one of the most common starting points for IT departments cutting their teeth is systems log file analysis
But not just about Hadoop. Players like EMC, HDS, HP, IBM, Oracle and others all have offerings in the Big Data arenaZooming out, here are some of the common categories (there are many ways of categorising, this is more of a stack view)Scale out storage can be used for large traditional warehouses as well as big dataDistributed analytics can run on proprietary grids as well as things like HadoopThe point being that no element of Big data is exclusiveIndeed there is nothing exclusive about big data at all – it will live alongside existing solutions (see next slide_
Has advantages that derivative insights are a lot more accessible when extracted from Big Data environment and placed in a traditional environment – skills sets/familiarity, toolset availability, well established integration and propagation mechanismsm,etc