• No results found

Quan%ta%ve Clustering of Cloud Compu%ng Projects for Be9er Standards Profiling

N/A
N/A
Protected

Academic year: 2021

Share "Quan%ta%ve Clustering of Cloud Compu%ng Projects for Be9er Standards Profiling"

Copied!
20
0
0

Loading.... (view fulltext now)

Full text

(1)

Quan%ta%ve  Clustering  of  Cloud  

Compu%ng  Projects  for  Be9er  

(2)

Neil  Caithness

 University  of  Oxford  

Michel  Drescher  European  Grid  Infrastructure  

Peter  Deussen

 Fraunhofer  FOKUS  

(3)

3  

Background

 

We  support  the  European  Commission’s  

vision  of  a  digital  single  market.  

Standardisa%on  is  perceived  as  a  strong  

enabler.  

We  support  over  70  EC-­‐funded  projects  

who  generally  all  have  standardisa%on  and  

interoperability  as  an  objec%ve.    

How  can  we  help  them?  

We  provide  a  repeatable  methodology  for  

cloud  standards  profiling.    

Contribu%ng  to  the  standards  landscape:  

IEEE  P2301,  ETSI  CSC2,  ISO/IEC  JTC1/SC38    

 

(4)

Set  of  common  standards  profiles  based  on  

clustering  of  ini%a%ves  on  the  NIST  cloud  

defini%on  characteris%cs.  

Increasing  awareness  of  security  cer%fica%on  

and  legal  issues  with  prac%cal  

recommenda%ons.  

Set  of  prac%cal  tools  to  support  cloud  

adop%on  for  SMEs  and  Public  

Administra%ons.  

(5)

5  

           Workflow  

Project  

Engagement  

 Project  selec%on  

 Use-­‐case  collec%on  

 Characteris%cs  scoring  

Data    

Analysis  

 Principal  Components  Analysis  (PCA)  

 Biplot  and  numerical  representa%on  

 Heirarchical  clustering  

Interpreta%on  

 Cluster  data  quality  

 Func%onal  vs.  non-­‐func%onal  characteris%cs  

Standards  

Profiles  

 Review  and  condense  project  use  cases  

 Cloud  standards  service  models  

 Review  standards  for  profiling  

Quan%ta%ve  

Methodology  

di

ss

em

in

a%

on

   &

   i

te

ra%

on

 

(6)

Defining  characteris%cs  

of  cloud  compu%ng  

NIST  special  Publica%on  800-­‐145  [NIST-­‐800-­‐145]

 

Essen%al  Characteris%cs  

 On-­‐demand  self  service  

 Broad  network  access  

 Resource  pooling  

 Rapid  elas%city  

 Measured  service  

Common  Characteris%cs  

 Massive  Scale  

 Homogeneity  

 Virtualiza%on  

 Low  Cost  Somware  

 Resilient  Compu%ng  

 Geographic  Distribu%on  

 Service  Orienta%on  

(7)
(8)

PCA  

Biplot  

(9)

9  

PCA  

Biplot  

How  many  components  to  keep?  

Scree  plot  &  Kaiser-­‐Gu9man  criterion  

(10)

Projec%on  (heat  map)  

Numerical  representa%on  of  the  biplot  

(11)

11  

Signal  

µ

   

Noise  

σ  

SNR  

(12)

Clustering  

(13)

13  

1  

2  

3  

(14)
(15)
(16)
(17)
(18)
(19)

19  

Workflow  

Project  

Engagement  

 Project  selec%on  

 Use-­‐case  collec%on  

 Characteris%cs  scoring  

Data    

Analysis  

 Principal  Components  Analysis  (PCA)  

 Biplot  and  numerical  representa%on  

 Clustering  

Interpreta%on  

 Cluster  data  quality  

 Func%onal  vs.  non-­‐func%onal  characteris%cs  

Standards  

Profiles  

 Review  and  condense  project  use  cases  

 Cloud  standards  service  models  

 Review  standards  for  profiling  

di

ss

em

in

a%

on

   &

   i

te

ra%

on

 

(20)

References

Related documents