BI Saas Cloud Strategies for Telcos

Document Sample
BI Saas Cloud Strategies for Telcos Powered By Docstoc
					                                               Brussels	
  
                                                21	
  –	
  22	
  	
  
                                               March	
  ‘13	
  




                                                    ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
BI	
  Software	
  as	
  a	
  Service	
  	
  
&	
  
Cloud	
  Strategies	
  for	
  Telcos	
  	
  
David	
  M	
  Walker	
  
Data	
  Management	
  &	
  Warehousing	
  

                                                                             1	
  
                                                                                                                  Brussels	
  



Cloud	
  Computing	
  (Since	
  1970)	
  
                                                                                                                   21	
  –	
  22	
  	
  
                                                                                                                  March	
  ‘13	
  




                                                                                                                       ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                                                     In	
  1970	
  a	
  service	
  provider	
  
                                                                           would	
  have	
  owned	
  a	
  
                                                                      mainframe	
  and	
  provided	
  
                                                                      clients	
  with	
  a	
  connecFon	
  
                                                                          that	
  allowed	
  them	
  to	
  
                                                                      interact	
  with	
  the	
  system.	
  
                                                                       Many	
  companies	
  would	
  
                                                                     have	
  accessed	
  this	
  system	
  
                                                                       sharing	
  the	
  resource	
  on	
  
                                                                                    demand	
  
                                                                	
  
  The	
  terminology	
  has	
  changed	
  and	
  the	
  technology	
  has	
  been	
  updated	
  
                     but	
  the	
  funcFonality	
  remains	
  the	
  same.	
  	
  
  “Cloud	
  compuFng	
  is	
  the	
  use	
  of	
  compuFng	
  resources	
  (hardware	
  and	
  
 soMware)	
  that	
  are	
  delivered	
  as	
  a	
  service	
  over	
  a	
  network	
  (typically	
  the	
  
   Internet).	
  Cloud	
  compuFng	
  entrusts	
  remote	
  services	
  with	
  a	
  user's	
                                                   2	
  
                data,	
  soMware	
  and	
  computaFon.”	
  -­‐	
  Wikipedia	
  
                                         Brussels	
  



Cloud	
  Computing	
  Architecture	
  
                                          21	
  –	
  22	
  	
  
                                         March	
  ‘13	
  




                                              ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                                                       3	
  
                                                  Brussels	
  



The	
  ‘as	
  a	
  Service’	
  Stack	
  
                                                   21	
  –	
  22	
  	
  
                                                  March	
  ‘13	
  




                                                       ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  The	
  more	
  standard	
  the	
  
   business	
  applicaFon	
  the	
  
   higher	
  up	
  the	
  stack	
  it	
  is	
  
   possible	
  to	
  go	
  
•  Most	
  organisaFons	
  can	
  
   do	
  e-­‐mail,	
  ERP,	
  CRM	
  as	
  
   SaaS	
  
•  There	
  is	
  also	
  NaaS	
  –	
  
   Network	
  as	
  a	
  Service	
  
   that	
  provides	
  the	
  
   connecFvity	
  to	
  the	
  
   cloud	
  &	
  between	
  clouds	
  
                                                                                4	
  
                                                                                       Brussels	
  



Cloud	
  Models	
  
                                                                                        21	
  –	
  22	
  	
  
                                                                                       March	
  ‘13	
  




•  Public	
  




                                                                                            ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
    •  Services	
  provided	
  by	
  a	
  service	
  provider,	
  usually	
  
       offered	
  on	
  a	
  pay	
  per	
  use	
  model	
  measured	
  by	
  CPU,	
  
       Storage,	
  Network	
  Bandwidth,	
  AcFve	
  Seats,	
  etc.	
  
•  Private	
  
    •  Services	
  built	
  by	
  an	
  organisaFon	
  for	
  it’s	
  own	
  use	
  
       but	
  using	
  the	
  technologies	
  and	
  techniques	
  of	
  
       public	
  cloud	
  compuFng	
  –	
  really	
  only	
  available	
  to	
  
       large	
  corporates	
  and	
  governments	
  
•  Community	
  
    •  Services	
  built	
  by	
  a	
  group	
  of	
  organisaFons	
  who	
  use	
  
       common	
  resources	
  or	
  share	
  common	
  concerns	
  
       (security,	
  compliance,	
  etc.)	
  
•  Hybrid	
  
    •  CombinaFon	
  of	
  Public	
  and	
  Private/Community	
  
       environment	
  using	
  allowing	
  commodity	
  
       components	
  where	
  possible	
  and	
  bespoke	
  cloud	
  
       services	
  where	
  necessary	
                                                                              5	
  
                                                                                                      Brussels	
  



Public	
  vs.	
  Private	
  Cloud	
  
                                                                                                       21	
  –	
  22	
  	
  
                                                                                                      March	
  ‘13	
  




                                                                                                           ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Public	
  cloud	
  has	
  an	
  iniFal	
  (very)	
  low	
  cost	
  with	
  predictable	
  
   long-­‐term	
  cost	
  –	
  although	
  the	
  long-­‐term	
  cost	
  is	
  oMen	
  
   considered	
  quite	
  high	
  –	
  private	
  cloud	
  is	
  costly	
  to	
  setup	
  and	
  
   long-­‐term	
  cost	
  is	
  unpredictable	
  
•  Scale-­‐up	
  and	
  scale-­‐out	
  are	
  easier	
  on	
  public	
  cloud	
  as	
  the	
  
   service	
  provider	
  is	
  responsible	
  for	
  ensuring	
  resources	
  are	
  
   available	
  at	
  a	
  pre-­‐determined	
  cost	
  
•  Security	
  &	
  privacy	
  of	
  data	
  on	
  public	
  cloud	
  will	
  always	
  be	
  a	
  
   concern.	
  There	
  are	
  addiFonal	
  concerns	
  over	
  data	
  locaFon	
  –	
  
   can	
  data	
  be	
  moved	
  between	
  states?	
  Can	
  rouge	
  states	
  host	
  
   their	
  applicaFons	
  in	
  the	
  same	
  hosted	
  environment	
  as	
  your	
  
   company	
  servers?	
  etc.	
  
                                                                                                                                    6	
  
                                                                                                                         Brussels	
  



Elastic	
  Services	
  
                                                                                                                          21	
  –	
  22	
  	
  
                                                                                                                         March	
  ‘13	
  




                                                                                                                              ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  In	
  order	
  to	
  get	
  more	
  resource	
  cloud	
  based	
  services	
  have	
  ways	
  to	
  
   enable	
  organisaFons	
  to	
  get	
  more	
  physical	
  or	
  virtual	
  servers	
  quickly	
  
   and	
  easily	
  

•  ElasFc	
  services	
  allow	
  organisaFons	
  to	
  expand	
  or	
  contract	
  the	
  size	
  of	
  
   their	
  individual	
  servers	
  (CPU/Disk/Bandwidth)	
  based	
  on	
  need	
  
    •  This	
  is	
  not	
  usually	
  transparent	
  as	
  it	
  oMen	
  requires	
  a	
  reboot	
  
    •  ElasFc	
  disk	
  space	
  is	
  usually	
  slower	
  than	
  dedicated	
  disks	
  

•  Choice	
  of	
  database	
  technology	
  will	
  affect	
  choices	
  
    •  Appliances	
  need	
  to	
  be	
  hosted	
  or	
  co-­‐located	
  
    •  Column-­‐store	
  databases	
  are	
  very	
  effecFve	
  in	
  elasFc	
  compute	
  as	
  the	
  
         they	
  use	
  less	
  resources	
  compared	
  to	
  the	
  raw	
  data	
  or	
  row	
  based	
  dbs	
  	
  
    •  Big	
  Data	
  technologies	
  favour	
  large	
  numbers	
  of	
  elasFc	
  compute	
                                                          7	
  
         technologies	
  but	
  this	
  can	
  get	
  expensive	
  
    	
  
                                                                                                                                          Brussels	
  



The	
  BI	
  Systems	
  Model	
  
                                                                                                                                           21	
  –	
  22	
  	
  
                                                                                                                                          March	
  ‘13	
  




                                                                                                                                               ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                   Database	
  




                                                                                     Database	
  
                                    Staging	
  	
  




                                                                                      Staging	
  	
  
                                                        Data	
                                                             ReporFng	
  
 Source	
                                                              ETL	
  	
                        Data	
  Mart	
  
               ETL	
  Server	
                        Warehouse	
                                                            Tools	
  
Systems	
                                                             Server	
                          Database	
  
                                                      Database	
                                                            Servers	
  




•  Most	
  BI/DWH	
  soluFons	
  do	
  &	
  will	
  look	
  something	
  like	
  this	
  
•  There	
  will	
  be	
  a	
  corporate	
  strategy	
  for	
  operaFonal	
  systems	
  
    •  Legacy	
  systems	
  will	
  be	
  in-­‐house	
  
    •  New	
  systems	
  may	
  be	
  in-­‐house	
  or	
  in	
  the	
  cloud	
  
    •  Even	
  if	
  an	
  organisaFon	
  is	
  commifed	
  to	
  a	
  fully	
  cloud	
  based	
  future	
  
       there	
  will	
  be	
  a	
  transiFonal	
  state	
  which	
  is	
  mixed	
  mode	
  
                                                                                                                                                                        8	
  
                                                                                                                                                Brussels	
  



Deployment	
  Model	
  1	
  
                                                                                                                                                 21	
  –	
  22	
  	
  
                                                                                                                                                March	
  ‘13	
  




                                                                                                                                                     ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                       Database	
  




                                                                                         Database	
  
                                        Staging	
  	
  




                                                                                          Staging	
  	
  
                                                            Data	
                                                               ReporFng	
  
  Source	
                                                                 ETL	
  	
                        Data	
  Mart	
  
                   ETL	
  Server	
                        Warehouse	
                                                              Tools	
  
 Systems	
                                                                Server	
                          Database	
  
                                                          Database	
                                                              Servers	
  



                                        In	
  	
  House	
                                                          In	
  The	
  Cloud	
  


•  ExisFng	
  Data	
  Warehouse	
  retained	
  in	
  house	
  
•  Data	
  Marts	
  &	
  ReporFng	
  Tools	
  migrated	
  to	
  the	
  cloud	
  
       •  ETL	
  Server	
  has	
  to	
  deliver	
  ‘ETL’	
  (i.e.	
  data	
  ready	
  for	
  direct	
  load)	
  
       •  ReporFng	
  via	
  web	
  interface	
  
       •  Desktop	
  tools	
  used	
  via	
  Remote	
  Desktop	
  
•  Favoured	
  as	
  a	
  first	
  step	
  into	
  the	
  cloud	
  	
  
                                                                                                                                                                              9	
  
       •  Experiments	
  in	
  public	
  cloud	
  and	
  for	
  extranet	
  usage	
  
	
  
                                                                                                                                          Brussels	
  



Deployment	
  Model	
  2	
  
                                                                                                                                           21	
  –	
  22	
  	
  
                                                                                                                                          March	
  ‘13	
  




                                                                                                                                               ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                   Database	
  




                                                                                     Database	
  
                                    Staging	
  	
  




                                                                                      Staging	
  	
  
                                                        Data	
                                                             ReporFng	
  
 Source	
                                                              ETL	
  	
                        Data	
  Mart	
  
               ETL	
  Server	
                        Warehouse	
                                                            Tools	
  
Systems	
                                                             Server	
                          Database	
  
                                                      Database	
                                                            Servers	
  



                   In	
  	
  House	
                                                       In	
  The	
  Cloud	
  


•  Deployment	
  with	
  two	
  ETL	
  servers	
  
    •  First	
  ETL	
  Server	
  loads	
  the	
  data	
  warehouse	
  
    •  Second	
  ETL	
  server	
  has	
  to	
  deliver	
  ‘ELT’	
  to	
  ensure	
  that	
  performance	
  
       is	
  sufficient	
  and	
  uses	
  a	
  pull	
  model	
  



                                                                                                                                                 10	
  
                                                                                                                                              Brussels	
  



Deployment	
  Model	
  3	
  
                                                                                                                                               21	
  –	
  22	
  	
  
                                                                                                                                              March	
  ‘13	
  




                                                                                                                                                   ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                       Database	
  




                                                                                         Database	
  
                                        Staging	
  	
  




                                                                                          Staging	
  	
  
                                                            Data	
                                                             ReporFng	
  
 Source	
                                                                  ETL	
  	
                        Data	
  Mart	
  
                   ETL	
  Server	
                        Warehouse	
                                                            Tools	
  
Systems	
                                                                 Server	
                          Database	
  
                                                          Database	
                                                            Servers	
  



       In	
  	
  House	
                                                      In	
  The	
  Cloud	
  


•  Deployment	
  with	
  two	
  ETL	
  servers	
  
    •  First	
  ETL	
  Server	
  loads	
  the	
  staging	
  area,	
  this	
  is	
  normally	
  a	
  literal	
  
         staging	
  area	
  and	
  therefore	
  is	
  just	
  ‘EL’	
  (no	
  ‘T’)	
  
    •  Second	
  ETL	
  server	
  has	
  to	
  deliver	
  ‘ELT’	
  to	
  ensure	
  that	
  performance	
  
         is	
  sufficient	
  and	
  uses	
  a	
  pull	
  model	
  
    	
  
                                                                                                                                                     11	
  
                                                                                                                                                       Brussels	
  



Deployment	
  Model	
  4	
  
                                                                                                                                                        21	
  –	
  22	
  	
  
                                                                                                                                                       March	
  ‘13	
  




                                                                                                                                                            ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                          Database	
  




                                                                                                  Database	
  
                                           Staging	
  	
  




                                                                                                   Staging	
  	
  
                                                               Data	
                                                                   ReporFng	
  
   Source	
                                                                         ETL	
  	
                        Data	
  Mart	
  
                      ETL	
  Server	
                        Warehouse	
                                                                  Tools	
  
  Systems	
                                                                        Server	
                          Database	
  
                                                             Database	
                                                                  Servers	
  



In	
  	
  House	
                                                            In	
  The	
  Cloud	
  


  •  Deployment	
  with	
  a	
  single	
  cloud	
  based	
  ETL	
  server	
  
        •  Normal	
  BI	
  operaFons	
  but	
  all	
  of	
  the	
  BI	
  is	
  in	
  the	
  cloud	
  
  •  More	
  oMen	
  used	
  with	
  Private	
  cloud	
  due	
  to	
  security	
  concerns	
  



                                                                                                                                                              12	
  
                                                                                                                                              Brussels	
  



Deployment	
  Model	
  5	
  
                                                                                                                                               21	
  –	
  22	
  	
  
                                                                                                                                              March	
  ‘13	
  




                                                                                                                                                   ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                  Database	
  




                                                                                         Database	
  
                                   Staging	
  	
  




                                                                                          Staging	
  	
  
                                                       Data	
                                                                  ReporFng	
  
 Source	
                                                                  ETL	
  	
                        Data	
  Mart	
  
              ETL	
  Server	
                        Warehouse	
                                                                 Tools	
  
Systems	
                                                                 Server	
                          Database	
  
                                                     Database	
                                                                 Servers	
  



                                                          In	
  The	
  Cloud	
  


•  Fully	
  cloud	
  based	
  environment	
  
•  No	
  issues	
  around	
  data	
  movement	
  volumes	
  
•  Security	
  concerns	
  will	
  have	
  been	
  addressed	
  because	
  all	
  
   corporate	
  data	
  is	
  in	
  the	
  cloud	
  already	
  


                                                                                                                                                     13	
  
                                                                                                                                                  Brussels	
  



Legacy	
  Deployment	
  Model	
  	
  
                                                                                                                                                   21	
  –	
  22	
  	
  
                                                                                                                                                  March	
  ‘13	
  




                                                                                                                                                       ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                    Database	
  




                                                                                             Database	
  
                                     Staging	
  	
  




                                                                                              Staging	
  	
  
                                                         Data	
                                                                    ReporFng	
  
 Source	
                                                                   ETL	
  	
                           Data	
  Mart	
  
                ETL	
  Server	
                        Warehouse	
                                                                   Tools	
  
Systems	
                                                                  Server	
                             Database	
  
                                                       Database	
                                                                   Servers	
  


In	
  The	
  
                                                                       In	
  	
  House	
  
Cloud	
  

•  Fully	
  cloud	
  based	
  environment	
  for	
  operaFonal	
  systems	
  
•  Data	
  being	
  dragged	
  back	
  into	
  legacy	
  BI	
  system	
  
•  Issues	
  with	
  migraFng	
  legacy	
  BI	
  technologies	
  to	
  the	
  cloud	
  



                                                                                                                                                         14	
  
                                                                                                                Brussels	
  



Major	
  Issues	
  
                                                                                                                 21	
  –	
  22	
  	
  
                                                                                                                March	
  ‘13	
  




                                                                                                                     ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Volume	
  of	
  data	
  over	
  the	
  network	
  
    •  Whichever	
  (distributed)	
  deployment	
  model	
  is	
  used	
  at	
  some	
  point	
  
       significant	
  amounts	
  of	
  data	
  will	
  have	
  to	
  moved	
  across	
  the	
  
       network	
  –	
  this	
  is	
  the	
  slowest	
  component	
  of	
  any	
  BI	
  soluFon	
  and	
  
       introduces	
  cost	
  and	
  latency	
  
•  Punching	
  a	
  hole	
  in	
  the	
  firewall	
  
    •  Many	
  organisaFons	
  will	
  be	
  concerned	
  about	
  deploying	
  an	
  ETL	
  
       tool	
  in	
  the	
  cloud	
  that	
  needs	
  fill	
  read	
  access	
  to	
  all	
  it’s	
  major	
  
       systems	
  –	
  deployment	
  models	
  that	
  restrict	
  this	
  access	
  tend	
  to	
  be	
  
       more	
  popular	
  
•  Compliance	
  
    •  Can	
  you	
  move	
  your	
  data	
  outside	
  your	
  state	
  and	
  any	
  other	
  range	
  
       of	
  regulaFons	
  	
  
    •  US	
  -­‐	
  FISMA,HIPAA,	
  SOX;	
  EU	
  –	
  Data	
  ProtecFon	
  DirecFve;	
  etc.	
                        15	
  
                                                                                                      Brussels	
  



Appliances	
  in	
  the	
  cloud	
  
                                                                                                       21	
  –	
  22	
  	
  
                                                                                                      March	
  ‘13	
  




                                                                                                           ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  All	
  physical	
  appliances	
  in	
  the	
  cloud	
  have	
  to	
  be	
  hosted	
  or	
  co-­‐
   located	
  
    •  Oracle	
  Exadata,	
  IBM	
  Netezza,	
  InfoBright	
  Infopliance	
  
    •  All	
  offer	
  some	
  managed	
  service,	
  implementaFon	
  with	
  preferred	
  
       partners	
  or	
  self-­‐build	
  cloud	
  opFons	
  
•  Virtual	
  Appliance	
  providers	
  allow	
  elasFc	
  services	
  to	
  be	
  used	
  as	
  
   well	
  as	
  being	
  hosted	
  or	
  co-­‐located	
  
    •  EMC	
  Greenplum,	
  ParAccel	
  
    •  These	
  technologies	
  have	
  community	
  ediFons	
  which	
  allow	
  try	
  
       before	
  you	
  buy	
  
    •  Again	
  all	
  offer	
  some	
  managed	
  service,	
  implementaFon	
  with	
  
       preferred	
  partners	
  or	
  self-­‐build	
  cloud	
  opFons	
  	
  	
  	
  
                                                                                                             16	
  
                                                                                                        Brussels	
  



Column	
  Store	
  In	
  the	
  Cloud	
  
                                                                                                         21	
  –	
  22	
  	
  
                                                                                                        March	
  ‘13	
  




                                                                                                             ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Column	
  Store	
  databases	
  well	
  suited	
  to	
  cloud	
  implementaFons	
  
•  UK	
  operator	
  data	
  being	
  loaded	
  into	
  a	
  column	
  store	
  database	
  
   •    2.6	
  Billion	
  call	
  data	
  records	
  (CDR)	
  per	
  day	
  
   •    114Gb	
  of	
  data	
  in	
  raw	
  files	
  including	
  reference	
  data	
  
   •    <	
  2.5	
  hours	
  to	
  transfer	
  to	
  server	
  and	
  load	
  data	
  
   •    Database	
  grows	
  by	
  7Gb	
  per	
  day	
  –	
  effecFve	
  compression	
  16.25:1	
  


•  Players	
  in	
  the	
  market	
  offering	
  services	
  in	
  this	
  area	
  include:	
  
   •  AcFan	
  Vectorwise*,	
  HP	
  VerFca,	
  InfoBright*,	
  MicrosoM	
  SQL	
  Server	
  
      2012,	
  SAP	
  Sybase	
  IQ	
  
   •  *These	
  technologies	
  have	
  community	
  ediFons	
  which	
  allow	
  try	
  
      before	
  you	
  buy	
  –	
  others	
  have	
  low	
  cost	
  PoC	
  opFons	
                            17	
  
   •  All	
  available	
  in	
  co-­‐locaFon,	
  hosted	
  and	
  elasFc	
  service	
  models	
  	
  
                                                                                                              Brussels	
  



Big	
  Data	
  in	
  the	
  Cloud	
  
                                                                                                               21	
  –	
  22	
  	
  
                                                                                                              March	
  ‘13	
  




                                                                                                                   ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  The	
  technologies	
  that	
  was	
  designed	
  for	
  use	
  in	
  the	
  cloud	
  from	
  
   the	
  beginning.	
  
•  Lots	
  of	
  emerging	
  technologies	
  and	
  companies	
  
   •  Current	
  ones	
  which	
  are	
  definitely	
  interesFng	
  include	
  Acunu,	
  
      Cassandra,	
  DataStax,	
  Cloudera,	
  HortonWorks,	
  MarkLogic,	
  
      Couchbase,	
  etc.	
  but	
  this	
  is	
  not	
  a	
  definiFve	
  list	
  	
  
   •  All	
  the	
  big	
  vendors	
  (IBM,	
  Oracle,	
  EMC,	
  SAP,	
  etc.)	
  also	
  have	
  big	
  
      data	
  offerings	
  
•  Many	
  of	
  these	
  technologies	
  being	
  deployed	
  besides	
  
   tradiFonal	
  database	
  technologies	
  
   •  Very	
  low	
  cost	
  as	
  the	
  soMware	
  is	
  free	
  and	
  by	
  definiFon	
  they	
  use	
  
      elasFc	
  compute	
  
                                                                                                                     18	
  
                                                                                                           Brussels	
  



BI	
  Tools	
  In	
  the	
  Cloud	
  
                                                                                                            21	
  –	
  22	
  	
  
                                                                                                           March	
  ‘13	
  




                                                                                                                ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Technical	
  deployment	
  of	
  ReporFng	
  and	
  Data	
  VisualisaFon	
  
   Tools	
  in	
  the	
  cloud	
  is	
  the	
  easiest	
  part	
  of	
  the	
  process	
  	
  
•  Each	
  product	
  has	
  a	
  defined	
  deployment	
  model	
  just	
  as	
  they	
  do	
  
   in	
  an	
  in-­‐house	
  data	
  centre	
  
   •  Obtain	
  the	
  servers	
  (co-­‐located,	
  hosted	
  or	
  elasFc)	
  and	
  install	
  the	
  
      soMware	
  and	
  connect	
  to	
  the	
  data	
  servers	
  
•  Some	
  hosFng	
  and	
  infrastructure	
  companies	
  will	
  provide	
  pre-­‐
   deployed	
  soMware	
  packages	
  

•  The	
  quesFon	
  is	
  does	
  your	
  soMware	
  vendor	
  understand	
  offer	
  
   an	
  effecFve	
  SaaS	
  licencing	
  model	
  –	
  most	
  maintain	
  their	
  
   exisFng	
  ‘per	
  seat’	
  licence	
  model	
  rather	
  than	
  pay	
  by	
  use	
  model	
                  19	
  
                               Brussels	
  



Big	
  Data	
  Landscape	
  
                                21	
  –	
  22	
  	
  
                               March	
  ‘13	
  




                                    ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                      20	
  
                                                                                                          Brussels	
  



Full	
  Stack	
  Cloud	
  Providers	
  
                                                                                                           21	
  –	
  22	
  	
  
                                                                                                          March	
  ‘13	
  




                                                                                                               ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Some	
  companies	
  have	
  the	
  depth	
  of	
  tools	
  (ETL,	
  Database,	
  
   ReporFng)	
  and	
  faciliFes	
  to	
  provide	
  a	
  full	
  BI	
  in	
  the	
  cloud	
  stack	
  

    •  IBM	
  
         •  Datastage,	
  Netezza,	
  Cognos	
  and	
  various	
  Big	
  Data	
  Technologies	
  
    •  MicrosoM	
  
         •  SSIS,	
  SQL	
  Server	
  2012,	
  ReporFng	
  Services	
  and	
  Hadoop	
  
    •  Oracle	
  
         •  Oracle	
  Technology	
  on	
  Demand	
  stack	
  (Exadata,	
  Discoverer,	
  etc.)	
  
    •  SAP	
  BI	
  On	
  Demand	
  
         •  Business	
  Objects	
  reporFng	
  and	
  ETL	
  with	
  SAP	
  HANA	
  

                                                                                                                 21	
  
•  In	
  pracFce	
  some	
  are	
  more	
  mature	
  than	
  others	
  in	
  this	
  area	
  
                                                                                                         Brussels	
  



The	
  Positive	
  Impact	
  of	
  Cloud	
  
                                                                                                          21	
  –	
  22	
  	
  
                                                                                                         March	
  ‘13	
  




                                                                                                              ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  The	
  provisioning	
  of	
  technology	
  should	
  be	
  faster	
  
•  The	
  cost	
  base	
  should	
  be	
  more	
  predictable	
  for	
  compute	
  
   resources	
  
•  The	
  cloud	
  is	
  an	
  excellent	
  way	
  to	
  trial	
  new	
  technologies	
  and	
  to	
  
   run	
  rapid	
  developments	
  in	
  a	
  Skunk	
  Works	
  approach	
  




                                                                                                                22	
  
                                                                                                         Brussels	
  



The	
  Negative	
  Impact	
  of	
  Cloud	
  
                                                                                                          21	
  –	
  22	
  	
  
                                                                                                         March	
  ‘13	
  




                                                                                                              ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Most	
  BI	
  operaFons	
  have	
  to	
  deal	
  with	
  an	
  IT	
  department	
  
   already	
  –	
  this	
  doesn’t	
  change	
  and	
  may	
  even	
  be	
  more	
  complex	
  	
  
•  Always	
  consider	
  the	
  cost	
  of	
  moving	
  large	
  amounts	
  of	
  data	
  
   from	
  an	
  in-­‐house	
  system	
  to	
  the	
  cloud	
  –	
  this	
  is	
  oMen	
  the	
  
   breaking	
  point	
  for	
  cloud	
  based.	
  
•  The	
  reverse	
  may	
  also	
  be	
  true	
  –	
  your	
  operaFonal	
  systems	
  may	
  
   be	
  put	
  in	
  the	
  cloud	
  before	
  you	
  have	
  BI	
  there,	
  in	
  which	
  case	
  
   you	
  will	
  be	
  paying	
  to	
  download	
  the	
  data	
  
•  There	
  are	
  many	
  more	
  operaFonal	
  issues	
  than	
  technical	
  issues	
  
   around	
  moving	
  service	
  provision	
  to	
  the	
  cloud	
  

                                                                                                                23	
  
                                                                                                                       Brussels	
  



What	
  doesn’t	
  change	
  
                                                                                                                        21	
  –	
  22	
  	
  
                                                                                                                       March	
  ‘13	
  




                                                                                                                            ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
•  Cloud	
  and	
  SaaS	
  are	
  just	
  the	
  latest	
  iteraFon	
  of	
  cost	
  management	
  
   concepts	
  (Timeshare,	
  Outsourcing,	
  etc.)	
  for	
  IT	
  
•  They	
  can	
  help	
  with	
  the	
  baseline	
  costs	
  of	
  a	
  BI	
  project	
  by	
  providing	
  
   and	
  managing	
  the	
  commodity	
  components	
  of	
  the	
  environment	
  

•  But	
  …	
  

•  Most	
  BI	
  projects	
  are	
  about	
  an	
  interacFon	
  between	
  the	
  business	
  and	
  
   IT	
  systems	
  
•  They	
  are	
  about	
  rapidly	
  delivering	
  informaFon	
  in	
  a	
  constantly	
  
   changing	
  environment	
  
•  <10%	
  of	
  the	
  TCO	
  over	
  10	
  years	
  is	
  in	
  infrastructure	
  –	
  the	
  rest	
  is	
  in	
  
   people,	
  development	
  of	
  ETL,	
  support,	
  change	
  	
                                                           24	
  
                                        Brussels	
  



Geek	
  &	
  Poke’s	
  View	
  …	
  
                                         21	
  –	
  22	
  	
  
                                        March	
  ‘13	
  




                                             ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
                                               25	
  

hfp://geekandpoke.typepad.com	
  	
  
                                                                                                         Brussels	
  



About	
  Us	
  
                                                                                                          21	
  –	
  22	
  	
  
                                                                                                         March	
  ‘13	
  




                                                                                                          13	
  November	
  
  Data	
  Management	
  &	
  Warehousing	
  is	
  a	
  UK	
  based	
  consultancy	
  




                                                                                                          2012	
  
that	
  has	
  been	
  delivering	
  successful	
  business	
  intelligence	
  and	
  data	
  
                        warehousing	
  soluFons	
  since	
  1995.	
  	
  




                                                                                                              ©	
  2012	
  Data	
  Management	
  &	
  Warehousing	
  
                                               	
  
  Our	
  consultants	
  have	
  worked	
  with	
  major	
  corporaFons	
  around	
  
 the	
  world	
  including	
  the	
  US,	
  Europe,	
  Africa	
  and	
  the	
  Middle	
  East.	
  	
  
                                               	
  
     We	
  have	
  worked	
  in	
  many	
  industry	
  sectors	
  such	
  as	
  telcos,	
  
    manufacturing,	
  retail,	
  financial	
  and	
  transport.	
  We	
  provide	
  
 governance	
  and	
  project	
  management	
  as	
  well	
  as	
  experFse	
  in	
  the	
  
                               leading	
  technologies.	
  	
  
                                                                                                               26	
  
                                                                                Brussels	
  



Contact	
  Us	
  
                                                                                 21	
  –	
  22	
  	
  
                                                                                March	
  ‘13	
  




                                                                                 13	
  November	
  
•  Data	
  Management	
  &	
  Warehousing	
  




                                                                                 2012	
  
   •  Website:	
  hfp://datamgmt.com	
  
   •  Telephone:	
  UK	
  +44	
  118	
  321	
  5930	
  




                                                                                     ©	
  2012	
  Data	
  Management	
  &	
  Warehousing	
  
   •  	
          	
   US	
  +1	
  716	
  406	
  3311	
  
•  David	
  Walker	
  
   •    E-­‐Mail:	
  davidw@datamgmt.com	
  
   •    Telephone:	
  +44	
  (0)	
  7990	
  594	
  372	
  
   •    Skype:	
  datamgmt	
  
   •    White	
  Papers:	
          	
  hfp://datamgmt.com/white-­‐papers	
  
   •    PresentaFons:	
             	
  hfp://datamgmt.com/presentaFons	
  
   •    Blog:           	
          	
  hfp://datamgmt.com/blog	
  
                                                                                      27	
  
                                               Brussels	
  
                                                21	
  –	
  22	
  	
  
                                               March	
  ‘13	
  




                                                    ©	
  2013	
  Data	
  Management	
  &	
  Warehousing	
   datamgmt.com	
  
BI	
  Software	
  as	
  a	
  Service	
  	
  
&	
  
Cloud	
  Strategies	
  for	
  Telcos	
  	
  
Thank	
  You	
  



                                                      28	
  

				
DOCUMENT INFO
Shared By:
Stats:
views:4
posted:3/22/2013
language:
pages:28
Description: A presentation to the ETIS Business Intelligence & Data Warehousing Working Group in Brussels 22-Mar-13 discussing what Saas & Cloud means and how they will affect BI in Telcos