• No results found

Shining Light on Dark Data


Academic year: 2021

Share "Shining Light on Dark Data"


Loading.... (view fulltext now)

Full text


Shining Light on

Dark Data

November, 2013


White Paper

Sponsored by Hitachi Data Systems

Shining Light on Dark Data

Benjamin S. Woo

November, 2013


Dark  Data  is  often  misunderstood.  It  is  considered  wasteful,  costly  and  valueless.  However,  in   reality,  it  is  often  quite  the  opposite.  Dark  Data  represents  untapped  value  and  opportunity.  In   order  to  extract  the  value  from  Dark  Data,  we  need  to  be  able  to  understand  better  what  is   captured  in  Dark  Data,  the  type  of  data  it  is,  and  categorize  it.    

In  this  White  Paper,  Neuralytix  looks  at  what  Dark  Data  is,  and  how  enterprises  can  generate   value  and  create  competitive  advantage  from  Dark  Data  with  the  assistance  of  HDS’s  Hitachi   Data  Discovery  Suite  (HDDS)  and  Hitachi  Content  Platform  (HCP)  

HDDS  is  storage  and  data  platform  independent  –  it  has  the  ability  to  index  nearly  500   different  data  types,  in  over  50  different  languages.  This  capability  exceeds  any  that  the   human  mind  can  comprehend  or  process.  HDDS  also  enables  the  ability  to  search  for  data  that   may  have  been  buried  in  archives.  It  can  also  enable  discovery,  information  about  how  data   relates  to  other  data,  to  generate  new  awareness.  

The  Hitachi  Content  Platform  (HCP)  stores,  shares,  synchronizes,  protects,  preserves,  analyzes   and  retrieves  file  data  from  a  single  system,  and  extend  the  capabilities  of  traditional  file   storage  solutions.  It  has  a  policy  engine  that  allows  the  automation  of  day-­‐to-­‐day  IT  

operations  such  as  data  protection  and  readily  evolves  to  changes  in  scale,  scope,  applications,   and  storage  and  server  technologies  over  the  life  of  data.  

With  the  combination  of  the  output  that  HDDS  provides  along  with  HCP,  industry  knowledge   and  business  experience,  enterprises  of  all  sizes  should  expect  significant  return  on  the   investment  in  a  data  discovery  solution.    


Table of Contents


Overview  ...  2

Table  of  Contents  ...  3

Current  Overview  ...  4

What  is  Dark  Data  ...  5

How  does  Dark  Data  get  generated  ...  5

Dark  Data  comes  in  many  forms  ...  5

Search  &  Discovery  ...  6

Search  ...  7

Discovery  ...  7

Search  &  Discovery  and  Dark  Data  ...  8

Neuraspective™  ...  9

Generating  Value  from  Dark  Data  ...  9

Data  retention  is  costly,  data  reuse  is  valuable  ...  9

Reusing  data  requires  coordination  and  consolidation  ...  10

Discover,  Search  and  Profit  ...  11

One  Search  for  All  Data  ...  11

Minimizing  Time  to  Insight  ...  12

Conclusion  ...  13

About  the  Author  ...  14

About  Neuralytix™  ...  14  


Current Overview

Enterprises  today  collect  tremendous  amounts  of  data.  That  data  is  preserved  for  a  number   of  standardized  uses  including,  but  not  limited  to,  reporting,  business  intelligence,  and   regulatory  compliance.  

There  are  also  other  reasons  for  data  preservation.  These  include  sentimental  and   historical  reasons.  Whatever  the  purpose  of  preserving  data  is,  enterprises  soon  realize   that  they  have  more  data  than  their  regular  processes  require  or  use.  

The  data  that  is  not  used  on  a  regular  basis  is  considered  “Dark  Data”.  This  data  seems   extraneous,  and  is  often  under-­‐  or  never  utilized.  This  data  is  collected  often  as  part  of   transaction,  and  most  likely  is  innocuous  or  referential  data.  

The  problem  is  that,  particularly  resulting  from  regulatory  compliance,  enterprises  are   being  required  to  keep  more  and  more  data  that  would  fall  under  the  definition  of  Dark  

Data.  Preserving  this  data  is  expensive.  Dark  Data  is  no  less  costly  to  preserve  than  non-­‐

dark  data.  Apart  from  the  cost  of  keeping  Dark  Data,  Dark  Data  can  also  be  a  long-­‐term   liability:  for  example,  if  the  enterprise  is  sued,  and  potentially  incriminating  data  is   subsequently  found  within  Dark  Data,  exposing  the  enterprise  to  tremendous  losses.  

Rather  than  taking  the  “ostrich”  approach  of  ignoring  the  existence  of  Dark  Data,  

enterprises  need  to  take  a  much  more  proactive  attitude  towards  Dark  Data.  Dark  Data  can   be  very  valuable.  In  keeping  with  the  example  above,  by  being  proactive  about  discovering   data  that  may  cause  a  significant  financial  liability,  the  same  enterprise  could  proactively   mine  its  own  Dark  Data,  internally  exposing  the  potential  liability,  and  proactively  

remedying  the  situation  with  its  customers.    

Not  only  can  such  measures  reduce  punitive  damages,  they  can  also  improve  customer   loyalty  and  build  respect  and  confidence  in  the  customer  base  towards  the  company.   Outside  of  these  cost-­‐mediating  benefits,  through  the  use  of  Big  Data  technologies,  Dark  

Data  can  actually  help  enterprises  to  proactively  generate  revenue  and  create  competitive  


This  White  Paper  looks  at  how  Hitachi  Data  Systems  addresses  these  challenges  and  helps   to  shed  light  (i.e.  generating  positive  enterprise  value)  on  Dark  Data.  

What is Dark Data

The  IT  industry  has  not  formally  defined  Dark  Data.  For  most,  the  concept  of  Dark  Data   represents  data  that  is  kept  because  an  organization  feels  that  the  destruction  of  such  data   may  prove  detrimental  in  some  way.  In  other  words,  we  could  define  Dark  Data  as  data  that   we  keep  but  do  not  really  know  why  we  keep  it!  

How does Dark Data get generated

Dark  Data  is  generated  in  an  infinite  number  of  ways.  It  could  be  data  that  is  collected  by  

transactional  systems  using  off-­‐the-­‐shelf  software,  but  has  no  direct  relationship  to  the   type  of  business  being  run.  It  could  be  log  files  that  are  archived  for  many  years,  simply   because  that  is  how  it  has  always  been  done.  It  could  also  be  derivative  data  left  over  from   other  processes.  

One  potential  example  of  Dark  Data  is  raw  survey  data  from  many  years  ago,  for  products   that  an  enterprise  no  longer  manufactures  or  distributes.  

One  of  the  most  common  way  that  Dark  Data  gets  created  is  a  result  of  versioning:  old   versions  of  documents  and  other  unstructured  data  that  get  kept  for  recovery  or  referential   purposes.  

Dark Data comes in many forms

As  the  examples  above  illustrate,  Dark  Data  can  be  structured,  semi-­‐structured,  or   completely  unstructured.  In  fact,  it  is  mostly  the  semi-­‐  and  fully-­‐unstructured  data  that   present  the  biggest  problems.    

Unstructured  data,  as  the  name  suggests,  have  no  prescribed  relationships  or  correlation  to   any  other  data.  They  are  not  supposed  to;  hence  the  concept  of  “unstructured”.  However,   the  lack  of  associations  and  connections  is  what  causes  the  Dark  Data  classification  in  the   first  place.  These  data  are  “orphaned.”  Without  specific  linkages  back  to  other  data,  their   preservation  or  destruction  can  generate  similar  liabilities  for  an  enterprise.  


Examples  of  unstructured  data  include  notes  made  by  doctors  on  a  patient’s  electronic   medical  record  (EMR).  The  notes  are  often  captured  in  a  text  field;  but  it  is  very  difficult  to   mine  through  this  data,  as  it  often  does  not  have  a  prescribed  relationship  to  any  other  field   other  than  the  patient’s  identity.  This  same  kind  of  example  can  extend  to  any  task  that   involves  some  form  of  human  intuition  or  note-­‐taking  (e.g.  notes  from  a  repairman).    

Search & Discovery

Accepting  the  existence  of  Dark  Data,  the  question  begs  to  be  asked,  what  can  one  do  with  

Dark  Data?  To  address  this  question,  two  technologies  have  found  great  prominence  in  the  

last  20  years:  search  and  discovery.  

Search  and  discovery  are  distinct  and  complementary  technologies.  With  search,  a  user’s   goal  or  expectation  is  to  find  an  answer  to  a  question.  With  discovery,  the  user  is  more   likely  to  seek  awareness,  rather  than  specific  answers.  

A  good  example  of  the  difference  would  be  in  legal  litigation.  One  party  may  conduct  legal   discovery  and  suggest  to  the  other  party  that  through  discovery,  they  found  incriminating   evidence  against  the  second  party.  Unlike  search,  in  which  the  outcome  is  precise,  the   outcome  of  the  discovery  could  simply  be  that  emails  exist  that  may  prove  the  second  party   liable  to  damages.  

At  this  point,  the  second  party  may  feel  sufficiently  threatened,  and  settle.    

Alternatively,  the  second  party  may  feel  that  the  first  party  is  bluffing,  and  that  the  outcome   of  their  discovery  (or  the  first  party’s  awareness  of  the  existence  of  the  emails)  is  

insufficient.  In  this  situation,  the  first  party  would  need  to  search  the  discovered  emails,  for   specific  evidence,  that  would  lead  them  to  answer  the  question  of  whether  the  second  party   has  wronged  the  first.    



As  noted  in  the  example  above,  a  search  seeks  to  find  specific  answers  to  a  question.  Public   search  engines  such  as  those  provided  by  Google,  Yahoo!  and  Bing  are  prime  examples  of   search  technology  at  work.  These  search  engines  enable  a  user  to  ask  the  search  providers   to  find  probable  answers  to  the  question  posed.  

The  search  engines  then  go  through  a  ranking  process  to  list  out  what  they  believe  to  be  the   most  appropriate  response,  but  nonetheless,  the  outcome  is  to  seek  a  satisfactory  direct   answer  to  a  posited  question.  


On  the  other  hand,  discovery  is  not  as  specific.  The  user’s  goal  is  generally  one  of  

awareness.  One  might  seek  to  discover  why  a  computer  system’s  processors  are  taxed  at   9:01am  on  a  Monday  morning.  An  obvious  “answer”  may  be  a  very  high  number  of  people   booting  up  their  computers  to  start  a  work  week.  But  in  discovery,  we  are  not  looking  for   specific  answers,  only  awareness.  

“Data  mining”  as  this  process  is  sometimes  referred,  can  be  augmented  with  Big  Data   technologies  to  proactively  generate  value  and  create  competitive  advantage.  

Big  Data  can  process  hundreds  of  machine-­‐generated  logs  and  find  patterns  or  correlations   between  the  increase  in  system  processor  usage  and  the  number  of  users  logging  on.  The   outcome  of  the  discovery  could  be  in  the  form  that  a  very  high  number  of  users  are  

performing  boot  sequences  at  9:01am  each  Monday  compared  to  any  other  time  of  the  day,   and  day  of  the  week.  

The  difference  is  very  fine.  Nevertheless,  the  outcome  of  the  discovery  does  not  specifically   identify  workers  starting  their  computers  at  the  head  of  the  workweek.  It  simply  suggests  a   very  high  number  of  boot  sequences  taking  place  at  9:01am  on  Mondays  compared  to  other   times  and  days.  

It  is  up  to  an  interpreter  of  the  data  to  make  sense  of  the  outcome  of  the  discovery.  As  such,   discovery  is  all  about  an  awareness  of  a  situation  in  order  to  allow  intelligent  


Since  computers  can  process  much  more  data  than  the  human  mind  can  handle,  computers   can  take  vast  amounts  of  seemingly  disparate  data  and  create  linkages,  patterns,  or  

relationships  that  are  beyond  the  capabilities  of  the  human  mind.  These  types  of   transactions  are  prime  examples  of  Big  Data  at  work.  

These  discoveries  help  individuals  to  take  an  immense  amount  of  data  and  produce  

information  in  a  human-­‐consumable  quantity.  This  information  then  helps  humans  to  make   better  decisions  more  quickly,  and  most  importantly  in  a  more  informed  manner.  

Search & Discovery and Dark Data

By  using  varying  combinations  and  permutations  of  search  and  discovery  technologies;   enterprises  are  able  to  generate  value  from  Dark  Data.  Computers  can  do  so  with  an   efficiency  that  goes  beyond  any  group  of  people.  They  can  also  deal  with  data  that  would   otherwise  have  been  forgotten  or  ignored.    

Today,  the  combination  of  all  data  (dark  or  otherwise)  and  the  use  of  search  and  discovery   technologies  form  the  basis  of  what  we  call  Big  Data.  Already,  Big  Data  has  been  

demonstrated  to  help  enterprises:  

• Improve  output  efficiency;   • Improve  customer  targeting;  

• Personalize  offerings  for  individual  customers;   • Improve  customer  service;  and  

• Improve  revenue  and  profits.  

The  benefits  of  Big  Data  can  range  from  saving  a  few  thousand  dollars;  to  improving  profit   by  millions  or  billions  of  dollars;  to  stopping  medical  pandemics  from  occurring;  to  saving   lives  and  families  in  underprivileged  nations.  



Generating Value from Dark Data

Data  overall  is  like  currency.  Not  money,  but  currency.  Think  of  each  piece  of  data  you   having  as  a  $1  bill.  There  are  many  ways  of  protecting  it,  the  most  popular  of  which  is  to   put  it  into  a  bank.    

But  what  about  the  change  you  receive,  that  you  pile  up  in  a  jar?  What  about  the  lost  coins   in  your  couch?  For  many,  after  a  while,  these  innocuous  and  relatively  immaterial  coins  can   add  up.    After  several  months,  many  may  find  hundreds  or  even  thousands  of  dollars  

collected  over  time.  Now,  include  the  odd  $1  or  $5  bill  that  may  have  been  left  lying  around   as  a  result  of  change  that  you  received  and  unconsciously  or  hurriedly  shoved  into  your   pant  pocket.  These,  like  the  coins,  often  get  quickly  pulled  out  of  pockets  before  the  pants   get  put  into  the  laundry.  Again,  like  the  coins,  these  bills  can  collect  over  time.  

Imagine  what  this  could  buy?  In  fact,  if  invested  wisely,  what  could  have  become  of  the   previously  “immaterial”  amount?  Consider  the  value  that  is  lost  because  what  $1  could   have  purchased  one  year,  through  inflation,  is  worth  less  in  subsequent  years.  

Data  is  just  like  currency.  It  too  can  get  left  behind.  It  too  can  lose  value  over  time.  Except,   unlike  coins  stuck  in  a  sofa,  or  change  on  a  dresser,  enterprises  have  a  lot  of  “loose”  or  dark   data  lying  around.  

Regulatory  compliance  compounds  this  further.  Now,  there  is  data  that  the  enterprise  may   not  want,  yet  is  forced  to  retain.  

Data retention is costly, data reuse is valuable

Data  that  is  stored  with  no  immediate  use  will  lose  value  over  time.  This  is  costly  to  an   enterprise.  Data  that  is  preserved  or  archived  simply  for  the  purpose  of  being  available  is   also  costly.  There  is  the  cost  of  the  storage  media,  the  maintenance,  power  and  cooling,  the   real  estate  cost  of  the  storage  system  and  many  other  considerations  that  add  up  to  a   substantial  amount  of  investment  and  support  required  to  retain  data.  


So  instead  of  simply  retaining  data,  enterprises  need  to  look  at  reusing  data.  When  data  is   able  to  be  reused,  its  value  can  be  sustained.    

Reusing data requires coordination and consolidation

A  big  challenge  for  most  enterprises  is  that  the  data  that  can  be  reused,  or  should  be   reused,  are  often  found  in  disparate  locations,  or  silos.  The  applications  that  generate  and   maintain  the  data  may  even  be  owned  by  separate  organizations  inside  the  enterprise.  For   example,  call  center  data  may  be  owned  by  the  support  organization  through  their  function   specific  software.  Conversely,  sales  and  marketing  data  that  reside  in  a  customer  

relationship  management  (CRM)  system  may  not  be  available  (or  considered  useful)  to  the   research  and  development  (R&D)  department,  and  as  such,  the  R&D  department  have  no   purview  or  access  into  the  CRM  data.  

Apart  from  the  internal  politics  associated  with  data  ownership,  are  issues  such  as  data   coherency  and  duplicity  across  these  organization.  In  some  cases,  similar  data  may  be   conflicting  in  nature  due  to  the  fact  that  each  organization  has  collected  the  data  from  a   different  viewpoint  or  perspective.    

Adding  to  all  of  this  are  issues  related  to  corporate  and  regulatory  compliance  and   governance,  the  equitable  application  of  security,  retention  and  other  corporate  policies   across  each  and  every  silo  of  data.  The  “quality”  of  the  data  plays  a  major  role  in  an  

enterprise’s  ability  to  reuse  data  and  obtain  the  necessary  insight  and  intelligence  that  will   generate  strategic  value  to  the  enterprise.  

HDDS  helps  to  coordinate  and  consolidate  access  to  multiple  data  sources.  The  use  of  the   word  access  is  deliberate.  HDDS  avoids  the  costly  replication  of  data  that  would  only  lead  to   the  extra  costs  associated  with  the  additional  storage  capacity  and  management  of  that   capacity.  

Instead,  HDDS  deploys  a  scale-­‐out  indexing  technology  that  securely  processes  petabytes   (PBs)  of  data  to  facilitate  search  and  discovery  of  disparate  storage  systems  and  data  types.    


In  today’s  global  economy,  HDDS  has  been  enhanced  by  providing  the  ability  to  index  454   different  data  types  in  56  different  languages  including  character-­‐based  languages  (like   those  found  in  Asia).  

Discover, Search and Profit

The  result  of  being  able  to  discover  and  search  across  so  many  formats  and  languages  is   that  queries  can  be  generated  to  quickly  generate  informed  decision  support  systems.    

Often,  from  a  discovery  perspective,  it  can  help  not  only  to  prove  a  hypothesis,  but  by   leveraging  HDDS,  it  can  also  quickly  disprove  hypotheses.  This  is  critical,  as  it  allows   enterprises  to  dispense  with  ideas  that  may  not  yield  a  reasonable  return.  Being  able  to   determine  these  situations  quickly  is  one  way  enterprises  can  be  more  agile,  while  being   able  to  avoid  unnecessary  costs  at  the  same  time.  

In  the  same  vein  as  Hitachi  Data  Systems’  (HDS’s)  approach  to  data  storage,  whereupon,   the  philosophy  is  “one  platform  for  all  data”,  HDDS  allows  enterprises  to  have  “one  search  

for  all  data”.    

By  using  a  consolidated  and  universal  interface,  business  users  across  organizational   structures  can  collaborate  and  generate  a  true,  enterprise-­‐wide,  360o  view  that  allows  

decisions  to  be  made  for  the  greater  good  of  the  enterprise  rather  than  simply  for  the   benefit  of  one  organization,  and  perhaps  at  the  cost  of  another.  

One Search for All Data

HDDS  can  provide  “one  search  for  all  data”.  HDDS  can  capture  metadata  from  known   formats  including  X-­‐Rays,  EXIF  data  from  digital  images,  the  context  of  an  email,  as  well  as   metadata  from  popular  office  productivity  suits  (such  as  Microsoft®  Word,  Microsoft  Excel®  

and  Microsoft  PowerPoint®),  etc.  It  can  capture  all  types  of  metadata  from  almost  any  data  

object  or  file  type.  

For  some  users,  data  needs  organization  and  context  from  the  beginning.  Often,  the  type  of   data  that  requires  this  is  rich  media  data  –  such  as  digital  images  and  videos.  For  example,   in  the  healthcare  vertical,  this  might  include  MRI  or  X-­‐Ray  images;  in  the  media  and  


to  be  unstructured.  But  in  order  to  gain  perspective  and  context,  these  data  need  to  be   organized  in  a  way  that  goes  beyond  a  simple  filename.  Metadata  (data  that  describes  other   data)  is  necessary.  

For  an  X-­‐Ray  image,  data  that  might  be  collected  in  metadata  may  be  the  body  part  being  X-­‐ rayed,  the  patient’s  identifier,  gender,  age,  and  perhaps  codes  that  would  identify  the  cause   of  the  x-­‐ray  being  required  in  the  first  place.  But  the  question  is  how  do  you  associate  that   type  of  information?  Most  file  systems  do  not  provide  the  ability  to  have  user-­‐defined   metadata.  Most  file  system  metadata  is  restricted  to  user,  group,  read-­‐only,  creation  and   modification  dates.  

Minimizing Time to Insight

When  it  comes  to  search  and  discovery,  time  to  insight  is  one  of  the  key  performance   metrics.  Time  to  insight,  along  with  accuracy,  allows  business  users  to  make  well-­‐informed,   empirically  supported  decisions  quickly.  

Object  stores  like  Hitachi  Content  Platform  (HCP)  store,  share,  synchronize,  protect,  

preserve,  analyze  and  retrieve  file  data  from  a  single  system,  and  extend  the  capabilities  of   traditional  file  storage  solutions.  It  has  a  policy  engine  that  allows  the  automation  of  day-­‐ to-­‐day  IT  operations  such  as  data  protection  and  readily  evolves  to  changes  in  scale,  scope,   applications,  and  storage  and  server  technologies  over  the  life  of  data.  

HCP  multi-­‐tenancy  capabilities  allow  it  to  support  a  variety  of  simultaneous  workloads  and   the  advanced  metadata  capture,  update  and  search  capabilities  provide  intelligence  tools   and    bring  structure  to  unstructured  file  data.  This  intelligence  automates  and  facilitates   deeper  analysis  of  data.  With  the  HCP  Anywhere  secure  file  synchronization  and  sharing   option  and  Hitachi  Data  Ingestor  (HDI),  data  can  be  collected  and  accessed  from  mobile   devices  and  remote  offices  while  being  centrally  managed.  



Dark  Data  has  a  lot  of  potential  value.  It  is  just  waiting  for  users  to  take  advantage  of  this  

potential  through  search,  discovery  and  Big  Data.  Users  need  systems  that  can  help  index   and  manage  metadata  in  order  to  make  Dark  Data  valuable.  

With  the  multitude  of  different  data  objects  and  data  types,  HDDS  is  a  leading  solution  in   helping  enterprises  understand  what  data  enterprises  are  actually  storing,  and  assist   business  analysts  to  leverage  the  untapped  value  within  Dark  Data.  

Extending  HDDS  capabilities  with  the  policy  driven  HCP  platform  creates  a  “perfect  storm”   of  scalability,  manageability,  “searchability”,  and  “discoverability”  of  data  located  anywhere   in  an  organization.  

With  just  a  little  effort,  enterprises  of  all  sizes  will  find  Dark  Data  lurking  around  their   storage  systems  –  from  USB  thumb  drives  to  monolithic  storage  systems.  This  data,  for  the   most  part,  is  preserved  innocuously,  for  sentimental,  historic,  referential  or  regulatory   reasons.  Yet,  this  data  holds  the  potential  to  generate  new  enterprise  value,  and  create   competitive  advantage.  

However,  unless  this  data  is  properly  indexed,  so  that  intelligence  can  be  applied  to  it,  and   value  extracted  from  it,  through  search  and  discovery,  Dark  Data  is  simply  costly.    

Neuralytix  believes  that  enterprises  of  all  sizes  will  need  to  integrate  solutions  that  enable   data  discovery  into  their  datacenter,  or  risk  becoming  uncompetitive.  In  a  negatively   polarized  case,  enterprises  can  leverage  HDDS  for  loss  mitigation.  When  used  properly,  an   investment  in  HDDS  will  yield  significant  return  to  the  enterprise.    


About the Author

Benjamin  Woo  is  the  Managing  Director  of  New  York  based  Neuralytix,  Inc.  

During  the  course  of  his  career,  Mr.  Woo  has  advised  clients  whose  collective  market   capitalization  is  over  US$1  Trillion.  Prior  to  founding  Neuralytix,  Mr.  Woo  was  Program   Vice  President  of  IDC's  Worldwide  Storage  Systems  Research,  where  he  led  a  team  of   analysts  responsible  for  advising  clients  on  the  evolution  and  trends  related  to  data  storage   systems;  the  impact  storage  systems  have  on  adjacent  technologies  including  servers,   software,  cloud  and  virtualization;  and  best  practices  in  go-­‐to-­‐market  strategies  related  to   the  storage  industry.    

In  addition  to  authoring  thoughtful  and  provocative  insight  on  the  storage  industry,  Mr.   Woo  is  a  frequently  sought  speaker  at  industry  and  customer  events  worldwide  and  is   frequently  quoted  in  the  leading  business  and  technology  presses.  While  at  IDC,  Mr.  Woo   also  initiated  the  global  Big  Data  research.  Mr.  Woo  has  keynoted  Big  Data  conferences   worldwide,  advising  leading  and  emerging  vendors  across  the  technology  spectrum  on  how   they  can  exploit  the  Big  Data  opportunity.    

About Neuralytix™

Neuralytix  is  the  global  leader  in  IT  research  and  consulting  focused  on  advising  vendors   and  business  users  on  strategies  that  maximize  competitive  advantage  and  generate   enterprise  value.  

Visit  http://www.neuralytix.com  to  learn  more.  


©  Copyright  2013,  Neuralytix,  Inc.  All  rights  reserved.  Reproduction  is  forbidden  unless   authorized.    For  reprints,  web  rights,  and  consulting  services  please  contact  Neuralytix  via   email  at  sales@neuralytix.com.  HITACHI  is  a  trademark  or  registered  trademarks  of   Hitachi,  Ltd.    


Related documents

In this paper a feasibility study of brain MRI dataset classification, using ROIs which have been segmented either manually or using a superpixel based method in

For example, Michigan Virtual University’s (2002) standards for quality online courses , the peer review proforma developed by the Griffith Institute for Higher Education

More interesting is the epitaph of Q.Atilianus, praefectus militum cohortis Hispanorum, at Sufasar in the Chelif valley. The episode implies the existence of an

Adubofuor, Isaac Amoah, Pearl Boamah Agyekum(2016), Physico-chemical properties pf pumpkin fruit pulp and sensory evaluation of pumpkin pineapple juice blend,

A meticulous review of the ERP implementation and assimilation literature would suggest that the technology-organisation-environment (TOE) framework (Tornatzky and Fleischer, 1990)

The Associate Athletic Director for Academics explained that in academics they are “focused a lot on [the] student’s progress towards their majors,” but wishes she could