• No results found

E-BUSINESS SUITE IN THE AMAZON CLOUD

N/A
N/A
Protected

Academic year: 2021

Share "E-BUSINESS SUITE IN THE AMAZON CLOUD"

Copied!
8
0
0

Loading.... (view fulltext now)

Full text

(1)

E-­BUSINESS  SUITE  IN  THE  AMAZON  CLOUD      

By  Chuck  Edwards,   President,  Blue  Gecko,  Inc.  

About  Blue  Gecko    

Blue  Gecko  is  a  remote  database  administration  company  headquartered  in  Seattle,  WA.    Founded  in   2001,  Blue  Gecko,  recipient  of  the  2009  Inc  5000  award  for  the  fastest  growing  companies  in  the   United  States,  provides  seven  by  twenty-­‐four  remote  support  for  over  100  Oracle,  Oracle  E-­‐Business   Suite,  MySQL,  and  SQL  Server  customers  around  the  world.    

Executive  Summary    

As  the  buzz  around  Cloud  Computing  has  grown,  so  has  confusion  and  marketing  misinformation   over  what  Cloud  Computing  actually  is.    Because  of  this  confusion,  enterprise  computing  customers,   such  as  those  who  use  Oracle  E-­‐Business  Suite,  often  do  not  have  a  clear  understanding  of  how  they   can  leverage  the  advantages  of  true  Cloud  Computing.        

Amazon  Web  Services  (AWS),  a  pioneer  and  leader  in  the  Cloud  Computing  space,  has  built  a  robust   cloud  platform  used  by  thousands  of  people  and  organizations  all  over  the  world.    Since  its  inception   in  2006,  developers  and  startup  companies  have  flocked  to  AWS  for  flexible,  inexpensive,  pay-­‐as-­‐you-­‐ go  utility  computing  services.    Today,  AWS  has  matured  into  a  viable  platform  not  just  for  developers   and  startups,  but  for  enterprise  customers  as  well.    

This  paper  will  explore  the  definition  of  cloud  computing  with  a  particular  focus  on  Amazon  Web   Services,  including  details  for  a  few  of  their  most  popular  services.    It  will  then  turn  to  enterprise  use   cases  for  AWS;  specifically,  this  paper  will  focus  on  examples  where  Oracle  E-­‐Business  Suite  

customers  can  leverage  the  flexibility  and  efficiency  of  Amazon's  cloud.      

What  is  Cloud  Computing?    

The  definition  of  Cloud  Computing  has  become  somewhat  muddled  by  marketing  campaigns  and   misunderstanding.    To  arrive  at  an  accurate  definition  of  Cloud  Computing,  it's  effective  to   understand  what  Cloud  Computing  is  not:    

Virtualization  is  not  Cloud  Computing.        

Virtualization  is  a  technique  used  to  run  more  than  one  virtual  machine  (VM)  on  a  single,  physical   host.    Virtualization  is  a  common  technique  used  in  a  Cloud  Computing  environment,  but  just  because   a  VM  runs  on  VMware,  Xen,  or  Microsoft's  hypervisor  does  not  mean  it  is  running  in  the  cloud.    

Software  as  a  Service  (SaaS)  is  not  Cloud  Computing.        

SaaS  is  SaaS  -­‐  that  is  to  say,  consumers  of  software  that  requires  only  a  browser  like  Salesforce.com   are  not  "Cloud  Computing"  any  more  than  they  are  when  they  buy  something  on  Amazon.com.    Cloud   Computing  refers  not  to  what  the  computing  infrastructure  is  doing  (i.e.  the  software),  it  refers  to  the   computing  infrastructure  itself.    It  is  more  accurate  to  call  Cloud  Computing  IaaS,  or  Infrastructure  as   a  Service.    

(2)

Managed  Hosting  is  not  Cloud  Computing.        

Though  managed  hosting  removes  the  burden  of  managing  data  centers  and  physical  hardware  from   the  end  user,  it  is  still  not  truly  Cloud  Computing.    The  managed  hosting  customer  still  must  operate   within  traditional  restrictions  such  as  minimum  resource  purchase  requirements,  minimum  time   commitments,  and  setup  time  and  fees.    Also,  managed  hosting  is  almost  always  tied  to  a  geography;   computing  resources  cannot  simply  be  allocated  in  one  geographic  region  or  another  instantly  and   automatically.    

True  Cloud  Computing  is  a  utility.    

In  true  Cloud  Computing,  the  end  user  interacts  only  with  the  service,  is  not  required  to  know   anything  about  the  underlying  service  provision  mechanisms  (i.e.  total  obfuscation),  and  pays  only   for  what  they  use.    Like  electricity,  natural  gas,  or  even  some  telephone  services,  usage  and  fees  are   incremental  and  may  be  scaled  up  and  down  as  required.        

Only  industry  insiders,  the  very  informed,  or  the  unusually  curious  attach  detailed  knowledge  of   residential  electricity  generation  and  delivery  to  the  simple  act  of  turning  on  a  light  switch.    And  so  it   is  with  Cloud  Computing:    Servers  may  be  activated  and  decommissioned;  storage  may  be  allocated   or  removed;  security  features  may  be  activates  and  modified  all  via  simple  API  calls  (i.e.  the  light   switch)  with  no  knowledge  of  what  happens  after  service  is  requested.  

Benefits  of  Cloud  Computing    

Cloud  Computing  benefits  revolve  around  management  and  scale.    A  true  cloud  provider  obfuscates   hardware,  network,  data  center,  and  all  physical  computing  resources  from  the  end  user.    Much  like   using  a  telephone  requires  no  knowledge  of  telephony  or  telephone  infrastructure,  the  cloud   customer  simply  uses  computing  resources  without  worrying  about  the  physical  infrastructure.    It  is   not  the  responsibility  of  the  consumer  to  replace  parts,  buy  support,  manage  facilities,  or  run  cable.     The  cloud  provider  handles  all  physical  and  integration  tasks  and  rolls  the  cost  into  the  usage-­‐based   computing  charges.    

Cloud  Computing  also  has  the  advantage  of  incremental  scaling.    Traditional  data  center  hardware   and  computing  infrastructure  is  usually  purchased  to  accommodate  current  peak  load  with  at  least   some  room  for  growth.    When  the  current  infrastructure  can  no  longer  handle  peak  load,  additional   hardware  is  purchased  with  even  more  room  for  growth.    An  unfortunate  side  effect  of  traditional   hardware  is  that  it  does  not  easily  scale  down;  typically  hardware  investments  are  a  sunk  cost,  and   do  not  simply  goo  away  when  you  don't  need  them  anymore.    With  Cloud  Computing,  consumers   avoid  all  of  these  hassles.    They  may  increase  and  reduce  resource  consumption  as  needed,  paying   only  for  what  is  used.    

Amazon  Web  Services  (AWS)  Overview    

AWS  is  a  true  Cloud  Computing  resource.    All  AWS  services  may  be  accessed  via  a  simple  SOAP/XML   API  call.    Customer  accounts  are  not  charged  setup  or  decommission  fees,  and  any  service  may  be   used  for  exactly  as  long  as  required;  customers  pay  only  for  what  they  use.    

Because  access  to  AWS  is  API  driven,  customers  may  access  computing  resources  in  many  ways.    The   AWS  API  is  public,  and  may  be  used  by  anyone  familiar  with  SOAP/XML,  Perl,  Ruby,  Java,  or  .NET   programming.    Amazon  provides  a  library  of  shell  scripts  and  a  web-­‐based  console  as  well.     Consumers  may  also  develop  their  own  interfaces,  such  as  the  ElasticFox  plugin  for  Firefox.    

(3)

Regions  and  Availability  Zones    

AWS  physical  resources  are  grouped  into  regions  and  zones.    A  region  is  a  general  geographic  area.     As  this  paper  is  written,  there  are  3  regions:    Eastern  US  (Virginia),  Western  US  (Northern  California),   and  Ireland  (Dublin).    Regions  are  further  subdivided  into  availability  zones.    For  practical  discussion,   a  zone  may  be  thought  of  as  a  distinct  data  center.1    Availability  zones  contain  no  common  single  

points  of  failure,  and  are  designed  to  provide  regional  redundancy  for  many  web  services.     A  complete  list  of  AWS  services  may  be  found  at  http://aws.amazon.com.    For  the  purposes  of  this   paper,  we  will  focus  on  the  three  services  most  relevant  to  running  E-­‐Business  Suite  in  the  Amazon   Cloud:    EC2,  EBS,  and  S3.    

Elastic  Cloud  Compute  (EC2)    

EC2  is  a  web  service  that  provides  on-­‐demand,  incremental  access  to  compute  capacity  in  the  cloud.     Servers  called  "instances"  are  available  in  a  variety  of  sizes  from  a  small  server  with  1  CPU  core  and   1.7GB  of  memory  to  high  capacity  servers  with  8  cores  and  68.4GB  of  memory.    EC2  instances  are   billed  hourly  for  actual  use  only.    Pricing  ranges  from  $0.085  to  $2.88  for  a  Linux  instance,  depending   on  the  instance  size  and  region  in  which  it  is  launched.    

Instances  are  launched  from  stored  profiles  called  Amazon  Machine  Images,  or  AMIs.    An  AMI  can  be   built  from  an  existing  EC2  instance  or  any  running  x86  or  x86_64  operating  system;  customers  may   build  an  AMI  from  a  specific  configuration  -­‐  a  Linux  server  running  Oracle  E-­‐Business  Suite,  for   example  -­‐  then  launch  as  many  instances  of  that  AMI  as  desired.    

Like  all  AWS  components,  EC2  instances  may  be  controlled  via  APIs:    Starting,  stopping,  reading   instance  status,  are  all  performed  via  API  or  through  an  API-­‐driven  application  such  as  the  AWS  Web   Console.    

EC2  Security  Zones    

All  EC2  instances  are  launched  within  a  security  zone.    A  security  zone  is  functionally  equivalent  to  a   network  firewall.    Ports  may  be  opened,  closed,  and  restricted  to  an  IP  address  or  range  of  IP   addresses.    Single  or  multiple  machines  may  be  launched  inside  a  security  zone;  when  multiple   machines  share  a  security  zone,  they  share  the  security  zone  rules.    

Security  zones  are  included  with  EC2  and  generate  no  separate  charges.    

Virtual  Private  Cloud  (VPC)    

A  VPC  is  a  private  network  within  the  AWS  cloud  that  is  tied  to  the  customer  network  via  an  IPsec   VPN  tunnel.    VPC  allows  customers  who  want  EC2  instances  to  stay  completely  private  to  extend   their  own  IP  space  into  the  AWS  cloud,  and  prevent  traffic  from  entering  or  leaving  the  VPC  except  as   controlled  by  their  local  network.    Most  Oracle  E-­‐Business  Suite  customers  would  almost  certainly   use  a  VPC  to  restrict  access  to  the  applications.    

A  VPC  costs  about  $36  per  month.    

                                                                                                               

1  Strictly  speaking,  an  availability  zone  is  not  a  data  center.    Two  data  centers  in  close  proximity,  for  

(4)

Elastic  Block  Storage  (EBS)    

Amazon  Elastic  Block  Store  (EBS)  provides  block  level  storage  volumes  for  use  with  EC2  instances.     Amazon  Elastic  Block  Store  provides  highly  available,  highly  reliable  storage  volumes  that  can  be   attached  to  a  running  Amazon  EC2  instance  and  exposed  as  a  device  within  the  instance.    

EBS  devices  can  be  created  and  attached  to  instances  via  API,  and  may  range  in  size  from  1GB  to  1TB.     Because  the  storage  in  an  EBS  volume  is  allocated  upon  creation,  usage  fees  are  based  on  allocation   even  if  the  entire  volume  is  not  used;  however,  EBS  only  costs  $0.10  per  gigabyte  per  month.     Snapshots  may  be  taken  of  EBS  volumes.    Snapshots  are  stored  in  an  S3  bucket  (see  below)  and  can   be  used  as  a  volume  backup,  or  may  be  used  to  create  a  new  EBS  volume.    For  example,  a  snapshot  of   an  EBS  volume  containing  an  E-­‐Business  Suite  APPL_TOP  could  be  used  to  create  a  new,  identical   volume  for  cloning  with  two  API  commands  (create  snapshot,  create  volume  from  snapshot).    

Simple  Storage  Service  (S3)    

Amazon  S3  is  a  different  type  of  storage  than  EBS.    Where  EBS  is  always  attached  to  an  EC2  instance,   S3  can  be  used  to  store  and  retrieve  any  amount  of  data,  at  any  time,  from  anywhere  on  the  web.     Objects  (i.e.  files,  images,  etc.)  are  redundantly  stored  on  multiple  devices  across  multiple  zones  in  an   Amazon  S3  Region.    A  chunk  of  S3  storage  is  called  a  "bucket."  

Amazon  S3  provides  versioning,  which  may  be  used  to  preserve,  retrieve,  and  restore  every  version   of  every  object  stored  in  an  Amazon  S3  bucket.    Customers  can  easily  recover  from  both  unintended   user  actions  and  application  failures  by  using  versioning.    

S3  costs  $0.15  per  gigabyte  of  used  storage  per  month,  and  is  billed  based  on  actual  consumption;  S3   storage  may  not  be  pre-­‐allocated.    

Common  E-­Business  Suite  Use  Cases    

AWS  components  are  much  like  UNIX  or  Linux  shell  programs:    Each  program  is  a  small,  flexible  tool   that  works  well  on  its  own,  but  really  shines  when  used  creatively  in  conjunction  with  other  

programs.    Similarly,  AWS  components  are  impressive  on  their  own,  but  can  be  made  to  perform   amazing  tasks  when  used  together.    

AWS  components  are  therefore  intentionally  de-­‐coupled  to  allow  for  maximum  flexibility;  it  is  up  to   the  admin,  developer,  or  businessperson  to  determine  the  appropriate  combination  for  their  needs.     Though  Oracle  E-­‐Business  Suite  has  very  specific  host  configuration  requirements,  AWS  provides  a   flexible  enough  platform  to  add  substantial  value  to  any  E-­‐Business  Suite  implementation.    

Disaster  Recovery    

Many  smaller  E-­‐Business  Suite  shops  have  no  active  disaster  recovery  environment.    Backup  to  tape   is  fine  for  archiving  backup  copies,  but  tape  needs  to  be  restored  in  case  of  emergency.    If  there  is  no   remote  data  center  immediately  available,  it  may  be  weeks  before  a  production  system  is  up  again.   Amazon  Web  Services  provides  a  flexible,  valid  alternative  to  traditional  disaster  recovery  hosting.     An  E-­‐Business  Suite  DR  site  may  be  set  up  in  a  matter  of  days  via  the  following  general  procedure:     1.    Set  up  a  VPC  if  a  private  network  is  desired.    

(5)

2.    Set  up  the  EC2  hosts  for  the  database  and  application  hosts  using  the  desired  AMI  and  instance   sizes.  

3.    Set  up  a  standby  database  from  the  existing  primary  database  to  the  EC2  database  host.     4.    Copy  the  application  file  systems  from  the  primary  to  the  EC2  application  hosts,  and  establish   rsync  or  any  other  desired  mechanism  for  file  system  synchronization.    

When  this  process  is  complete,  a  fully  functional,  private  DR  site  will  be  running  in  the  Amazon  Cloud   for  a  fraction  of  the  cost  of  dedicated  hardware  or  traditional  DR  services.        

Because  AWS  costs  are  so  low,  the  most  expensive  consideration  is  Oracle  licensing.  If  budget  is   paramount,  here  are  a  couple  of  ways  to  eliminate  licensing  costs  while  still  maintaining  a  valid  DR   site:    

1.    For  the  database  host,  keep  the  database  software  up-­‐to-­‐date  on  the  EC2  host,  but  do  not  start  a   standby  database.    There  are  no  licensing  fees  for  installed  software,  only  running  software.    Ship   RMAN  backup  copies  and  archived  redo  logs  to  the  EC2  instances  via  the  network  or  Amazon's   import/export  service.    If  the  DR  site  is  needed,  a  restore  and  recovery  will  be  much  faster  if  all  the   backup  pieces  are  immediately  available.    

2.    Like  the  database  host,  sync  the  application  tier  to  the  application  hosts  without  starting  the   application  services.    There  is  no  charge  for  copying  the  software  from  one  host  to  another  as  long  as   it  isn't  used.    Most  Oracle  licenses  have  a  provision  for  DR  testing  a  certain  number  of  days  per  year   at  no  additional  cost;  talk  to  your  Oracle  licensing  vendor  for  details.    

Test  Environments    

Test  environments  are  one  of  the  best  AWS  use  cases  for  E-­‐Business  Suite.    Test  environments  often   do  not  need  to  be  the  same  size  or  topology  as  production;  even  shops  that  have  large,  complex   production  environments  with  RAC  and  load-­‐balanced  web  servers  can  use  single-­‐node  or  two-­‐node   test  environments  for  development  and  application  configuration  testing.  

Test  environments  are  created  via  the  E-­‐Business  Suite  cloning  process.    Much  like  setting  up  a  DR   environment,  test  environments  require  a  current  copy  of  the  database.    To  maximize  flexibility,  set   up  a  DR  environment  on  EC2,  and  then  use  current  standby  copies  and  snapshots  of  the  synced   application  file  systems  to  quickly  build  up-­‐to-­‐the-­‐minute  test  environments  as  needed.    

Once  again,  because  hardware  deployment  is  nothing  more  than  a  few  API  calls  or  clicks  on  a  web   console,  test  environments  may  be  created  more  quickly  and  in  a  completely  automated  fashion  (API   calls  to  create  a  host  are  a  straightforward  extension  of  any  cloning  procedure.)    Once  a  test  

environment's  use  has  expired  -­‐  or  even  on  weekends  and  nights  when  they  are  not  being  used  -­‐   simply  shut  them  down  to  avoid  even  the  inexpensive  EC2  usage  fees.  

Backups  and  Backup  Archiving    

Amazon  S3  is  an  inexpensive,  efficient,  and  in  many  ways  superior,  alternative  to  tape  for  E-­‐Business   Suite  backups.    As  detailed  earlier,  S3  storage  is  replicated  in  at  least  three  distinct  geographic   locations  within  seconds  of  transfer.    When  considering  an  archive  or  static  backup  solution,  very  few   alternatives  compare:    

(6)

1.    Tape  backup.    Tape  backup  requires  a  local  tape  copy.    Even  if  a  vaulting  service  is  used,  the  tape  is   exposed  to  damage,  loss,  or  theft  during  the  time  it  waits  in  the  library  for  pickup  and  during  

transport.    Once  the  tape  reaches  the  off-­‐site  vault,  it  exists  only  in  that  vault;  in  a  truly  catastrophic   disaster,  someone  must  be  available  for  the  tape  to  be  retrieved  from  the  vault  and  delivered  to  the   DR  site.    Multiple  tape  copies  to  different  locations  (or  even  to  maintain  a  local  copy)  increases  costs.     2.    Hardware  or  software  replication.    Replication  is  superior  to  tape  in  many  ways,  but  is  still   typically  performed  only  to  a  single  target.    Because  replication  hardware  is  expensive,  multiple   targets  are  typically  not  practical,  especially  for  a  smaller  shop.    Software  replication  solves  some  of   the  hardware  replication  cost  concerns;  however,  additional  hardware  of  some  kind  is  still  required   in  multiple  locations,  which  requires  maintenance  and  overhead.    Further,  data  replication  is  not  a   long-­‐term  vaulting  solution  -­‐  SANs  and  backup  appliances  have  limited  capacities  and  eventually  the   data  must  find  its  way  to  tape  or  other  media  if  long-­‐term  vaulting  is  required.    

Direct  backup  to  S3  is  a  superior  backup  solution  for  smaller  data  volumes,  and  combining  S3  with   replication  or  near-­‐line  disk  backup  is  less  expensive  and  more  flexible  than  almost  any  traditional  or   specialized  backup  solution.      

S3  costs  $0.15  per  gigabyte  of  used  storage  per  month,  and  is  billed  based  on  actual  consumption.     Tape  libraries  and  backup  appliances,  not  to  mention  the  data  centers  in  which  they  reside  all  have   purchase  and  support  costs,  often  in  the  tens  of  thousands  of  dollars,  whether  1MB  or  10TB  is  used.     Oracle  has  certified  and  supports  database  backup  to  S3.    This  means  normal  RMAN  backups  may   address  an  S3  bucket  as  a  backup  target  just  like  traditional  tape  or  disk.    If  backup  are  large  and   network  performance  is  a  concern,  consider  the  cost  of  a  large,  fast  network  pipe  and  the  low  cost  of   S3  storage  versus  the  huge  investment  in  backup  equipment,  equipment  maintenance,  and  vaulting   charges.    

For  the  E-­‐Business  Suite  file  system,  simple  tools  like  rsync  or  ssh  can  copy  files  to  an  S3  bucket.     Many  commercial  vendors,  such  as  Zmanda,  already  provide  an  S3  bucket  as  a  backup  target  option.     S3  backup  and  vaulting  may  not  be  the  right  solution  every  time.    It's  not  difficult  to  pick  an  extreme   data  volume,  peculiar  storage  or  recovery  requirement,  or  site  an  arbitrary  corporate  policy  to   invalidate  S3  as  a  backup  option;  however,  for  most  shops  and  recovery  requirements,  S3  represents   a  cost-­‐effective  and  efficient  alternative  to  traditional  backup  and  vaulting  methods.    

Common  Concerns    

Despite  many  of  AWS's  obvious  advantages,  many  potential  customers  are  still  put  off  by  the  nature   of  Cloud  Computing;  the  same  obfuscation  that  makes  AWS  so  simple  and  flexible  may  also  breed   mistrust.    Some  of  the  most  common  concerns  resolve  around  security,  performance,  and  vendor   support.    

Security    

AWS  has  completed  a  SAS70  Type  II  Audit,  which  is  almost  an  industry  requirement  for  any  provider   who  wishes  to  host  Oracle  E-­‐Business  Suite  for  a  publicly  traded  company.    AWS  is  aggressively   pursuing  additional  security  certifications  and  accreditations  to  demonstrate  the  security  of  their   infrastructure  and  services.    

AWS  uses  the  same  secured  facilities  as  the  Amazon.com  web  site.    Housed  in  Amazon-­‐controlled   data  centers  throughout  the  world,  Amazon  limits  knowledge  of  the  actual  location  of  these  data  

(7)

centers  to  only  required  personnel.    The  data  centers  are  secured  with  a  variety  of  physical  barriers   to  prevent  unauthorized  access.    

AWS  services  encrypt  API  and  user  traffic  by  default.    For  example,  certificates  are  required  to   manipulate  APIs;  EC2  security  zones  are  locked  down  by  default;  ssh  keys  are  required  to  access  EC2   instances  upon  launch  unless  explicitly  disabled.    Detailed  security  information  for  each  service  (EC2,   S3,  etc.)  may  be  found  in  the  document:    Overview  of  Security  Processes  white  paper.    

Performance    

A  major  concern  for  database  customers  is  I/O  performance.    EBS  volumes  are  attached  to  EC2   instances  via  a  single  network  interface  using  iSCSI.    I/O  performance  may  be  greatly  enhanced  by   using  multiple  EBS  volumes.    For  example,  LVM  can  stripe  a  volume  and  ASM  can  create  a  diskgroup   across  multiple  EBS  devices.    These  techniques  improve  I/O  performance,  but  do  not  eliminate  the   potential  bottleneck  of  the  single  network  interface.    

For  most  customers,  I/O  throughput  will  not  be  a  problem  (need  example  from  Jeremiah);  however,   if  extreme  I/O  throughput  is  required,  one  of  the  benefits  of  AWS  is  that  unlike  expensive  hardware   test  labs,  it  is  very  inexpensive  to  test  every  possible  scenario  on  AWS.    

Support  and  Licensing    

As  this  paper  is  written,  running  an  Oracle  Database  on  EC2  and  RMAN  backups  to  S3  are  certified   and  supported,  and  has  published  licensing  guidelines.    Oracle  E-­‐Business  Suite  is  not  yet  certified   nor  supported  on  AWS,  and  the  use  cases  in  this  paper  respect  that  position  by  avoiding  production   AWS/E-­‐Business  scenarios;  however,  it  is  important  to  remember  that  with  the  notable  exceptions  of   AWS  and  Oracle's  own  virtualization  platform,  Oracle  certifies  their  products  on  operating  systems,   not  hypervisors  or  hardware.    

If  a  customer  encounters  a  bug  and  logs  a  TAR  with  Oracle  Support,  Oracle  will  attempt  to  re-­‐create   the  bug  and  develop  a  patch.    If  Oracle  Support  is  unable  to  replicate  the  bug,  they  may  determine   that  the  issue  is  not  a  bug,  but  is  a  problem  with  a  hardware  component  below  the  operating  system,   e.g.  a  SAN  or  network  device.    They  will  then  refer  the  customer  to  the  hardware  vendor  for  

resolution.    

It  is  the  same  for  hypervisor  vendors  such  as  VMware,  Citrix  (Xen),  and  Microsoft.    Running  Oracle  E-­‐ Business  Suite  on  an  "unsupported"  hypervisor  does  not  mean  that  Oracle  Support  will  turn  you   away  at  the  door  any  more  than  they  will  reject  your  TAR  out  of  hand  for  running  Oracle  on  an   ancient,  smoking  piece  of  junk  you  found  hiding  in  a  hot,  damp  corner  of  your  system  administrator's   garage.    It  simply  means  that  if  you  encounter  a  bug  that  Oracle  cannot  reproduce,  even  on  a  

supported  operating  system,  they  may  send  you  and  your  issue  to  your  hardware  or  hypervisor   vendor  for  support.    

Should  you  run  a  production  E-­‐Business  Suite  environment  on  AWS?    That  depends  on  your  risk   tolerance  and  the  likelihood  of  finding  a  bug  no  one  has  ever  encountered  before.    In  no  way  does   Blue  Gecko  recommend  running  Oracle  on  an  unsupported  platform,  but  as  outlined  above,  the   notion  of  support  becomes  a  bit  fuzzy  below  the  operating  system  except  in  very  specific  

circumstances.    For  these  reasons  the  use  cases  in  this  document  outline  real-­‐world  situations  for   leveraging  AWS  across  non-­‐production  environments  to  gain  efficiency,  flexibility,  and  cost  savings.     One  might  argue  that  the  disaster  recovery  use  case  implies  production  use  at  some  point,  and  this   argument  would  be  true.    That  same  argumentative  person  may  next  wish  to  internally  debate  the  

(8)

merits  of  an  affordable,  perfectly  functional,  but  questionably  supported  DR  site  in  certain  edge-­‐case   scenarios  versus  no  DR  site  at  all.        

Conclusions    

Oracle  E-­‐Business  Suite  is  expensive  software  that  traditionally  requires  expensive  computing   infrastructure.    A  properly  managed  E-­‐Business  Suite  shop  will  clone  production  to  test  

environments  for  development,  patch  testing,  and  other  non-­‐production  activities.    If  the  role  of  E-­‐ Business  Suite  is  critical  to  the  enterprise  -­‐  and  who  considers  their  multi-­‐million  dollar  investment   in  ERP  or  manufacturing  infrastructure  non-­‐critical?  -­‐  then  the  business  probably  cannot  withstand   too  much  downtime,  making  a  DR  site  of  some  sort  appealing.    

However,  when  the  financial  and  logistical  burden  of  maintaining  non-­‐production  computing   infrastructure  becomes  to  great,  prudent  steps  like  test  and  DR  environments  are  glossed  over  or   skipped  entirely.    Test  environments  are  not  refreshed;  developers  and  project  managers  share   environments  that  don't  represent  production;  DR  is  left  to  tape  backups  and  a  "We'll  cross  that   bridge  when  we  come  to  it,"  recovery  plan.    

Amazon  Web  Services  is  moving  aggressively  into  the  enterprise  computing  space  with  a  heavy   emphasis  on  enterprise  partners  like  Oracle.    In  all  likelihood,  Oracle  E-­‐Business  Suite  will  soon  be   supported  on  AWS.    But  the  advantages  provided  by  AWS  as  a  true  cloud  provider  -­‐  incremental   scaling;  low,  pay-­‐as-­‐you-­‐go  fees;  no  infrastructure  maintenance  whatsoever;  VPC  security  -­‐  may  be   leveraged  in  many  practical  ways  by  E-­‐Business  Suite  customers  today.      

           

References

Related documents

•  Fusion Intelligence for E-Business Suite.. • 

Oracle Financials Cloud is designed to work with other Oracle Applications Cloud Services and your existing applications portfolio, such as Oracle E-Business Suite and

The Extension for Oracle E-Business Suite allows users to schedule and automate processes within Oracle E-Business Suite product line and integrate those processes with other

Web Access File Permissions Web Services and SOA Network Network Testing Network Auditing Change Control Firmware Patches Network Encryption Provider Shared Customer Database

– SaaS Public Cloud (i.e. Oracle E-Business Suite).. What are some of the issues with

Sharpen your friends and free thompson chain reference bible clearly states that he who talks back online threats that particular study time protection uses the game.. Talks back

At the same, we should inquire if concepts and models that have been developed in the second half of the 20th century in order to address social, ethical and political dimen- sions

Oracle E-Business Suite R12.1 Financial Management Certified Implementation Specialist: Receivables (1Z0-518) OR Oracle E-Business Suite Support Specialist Assessment Oracle