Database & Technology 1 _ Marcelle Kratchvil _ Why you should be storing unstructured data in oracle databases.pdf
1. Why you should be storing
unstructured data
in the Oracle database
Marcelle
Speaker
Name
Kratochvil
CTO
Pic/on
Speaker
Organisa/on
marcelle@pic/on.com
Speaker
Date
"Over 80% of data is unstructured and not managed
- Isn't it about time you decided to manage it?"
The most comprehensive Oracle applications & technology content under one roof
2. Inaugural International Unstructured Data with
Multimedia SIG
At OpenWorld – Sunday 9am (Moscone West)
To gather experts in the field to share ideas and to help those who are
learning about multimedia and any unstructured data be able to find
resources and share programs, papers, and ideas on anything in this field.
For more information go to : http://eternal-donut.blogspot.com/
If you can't make OpenWorld, register at the blog site.
The most comprehensive Oracle applications & technology content under one roof
3.
4. Why should I store unstructured data in the database?
This
is
not
the
right
ques/on
to
be
asked
It should be
What
is
your
jus/fica/on
for
not
storing
it
in
the
database?
The most comprehensive Oracle applications & technology content under one roof
5. What
is
Unstructured
Data
?
l Mul/media
l Photos
Structured = Relational
l Audio
Semi Structured = XML
l Video
l Documents
l Files
l Raw
Data
The most comprehensive Oracle applications & technology content under one roof
6. Who
needs
it
?
l Museums
l Educa/on
Image Warehouses
l Government
l Marke/ng
l Defence
Its now hard to find a business that doesn't need it
The most comprehensive Oracle applications & technology content under one roof
7. The
Rules
Change
l Rela/onal
is
a
well
structured
subset,
with
clearly
defined
rules
l XML
is
structured
but
has
variable
rules
l Mul/media
introduces
fuzzy
concepts,
normal
referen/al
rules
stop
making
sense
l Other
data
is
the
wild
west
The most comprehensive Oracle applications & technology content under one roof
8. Interface
l iPhone
and
iPad
changed
the
interface
l Augmented
reality
l More
intui/ve
l Simpler
is
beUer
l Users
are
demanding
this
simplicity
The most comprehensive Oracle applications & technology content under one roof
9. Searching
l Exact
search
doesn't
make
sense
(unless
rela/onal)
l Metadata
is
the
rela/onal
link
l Fuzzy
Searching
l Find
images
that
look
like
l Find
a
scene
with
this
actor
in
it
l Find
the
posi/on
in
the
audio
file
where
the
person
says
l Find
the
audio
file
with
this
bird
song
in
it
The most comprehensive Oracle applications & technology content under one roof
10. Exactness
l Does
Image
A
=
Image
B?
l At
what
point
are
they
different?
l What
if
a
byte
changes?
l What
if
there
is
a
subtle
color
change
l Checksums
?
l Enforcing
compliance
and
copyright
l What
about
two
images
of
the
same
scene?
The most comprehensive Oracle applications & technology content under one roof
11. Processing
l Transform
There is no “right” format
l DNG,
raw
to store an Image in
l Convert
l One
format
to
another
l Modify
l Crop,
watermark
l The
data
needs
exercising
The most comprehensive Oracle applications & technology content under one roof
12. So
why
Store
it?
l Manageability
l Security
l Backup/Recovery
l Extensibility
l Flexibility
The most comprehensive Oracle applications & technology content under one roof
13. Manageability
l Focus
is
Metadata
l Kept
in
Sync
l Metadata
to
External
Images
is
painful
l Try
changing
a
loca/on
l SQL
/
Object
access
l Well
understood
l PL/SQL
access
The most comprehensive Oracle applications & technology content under one roof
14. Security
l Row
level
protec/on
l Audi/ng
l Dynamic
Security
l Encryp/on
l Proven
The most comprehensive Oracle applications & technology content under one roof
15. Backup/Recovery
l Block
level
checking
and
recovery
l Unified
backup
rela/onal
+
objects
using
rman
l Simpler
restore
l Full
use
of
rman
features
The most comprehensive Oracle applications & technology content under one roof
16. Extensibility
l Indexing
l Object
Oriented
l Processing
and
Conversion
The most comprehensive Oracle applications & technology content under one roof
17. Flexibility
l Manage
metadata
l Process
en-‐masse
l Link
images
together
l Image
Warehouse
l RAC,
Streams,
Replica/on,
Data
Guard,
Flashback
The most comprehensive Oracle applications & technology content under one roof
18. And
the
concerns?
l Maturity
l Performance
l Database
Size
l Complexity
The most comprehensive Oracle applications & technology content under one roof
19. Maturity
l Oracle
is
way
ahead
of
the
other
vendors
l Not
all
features
are
mature
l Oracle
Mul/media
l RAC
l Exadata
l Rman
l Streams
l Gateways
The most comprehensive Oracle applications & technology content under one roof
20. Performance
l Loading
and
Retrieval
l Securefiles
is
very
fast
l Processing
l Java
and
Mul/media
do
not
mix
well
l You
need
a
GPU
to
process
mul/media
l VM's
do
not
process
mul/media
well
The most comprehensive Oracle applications & technology content under one roof
21. Database
Size
l Block
overhead
wastage
l Larger
block
size
needed
l Securefiles
addresses
this
The most comprehensive Oracle applications & technology content under one roof
22. Complexity
l Mul/media
is
hidden
l Its
not
tangible
like
a
file
l You
need
an
applica/on
to
get
to
it
l Solved
by
Oracle
FileSystem
(Linux
Only)
The most comprehensive Oracle applications & technology content under one roof
23. Finally
l There
are
more
benefits
to
store
all
your
data
in
the
database
l Protect
it
l Secure
it
l Add
value
to
it
The most comprehensive Oracle applications & technology content under one roof
24. For
more
informa/on
l marcelle@pic/on.com
l hUp://eternal-‐donut.blogspot.com/
The most comprehensive Oracle applications & technology content under one roof