Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
The	
  Square	
  Kilometre	
  Array	
  (SKA)	
  
Next	
  Genera8on	
  radio	
  telescope	
  –	
  compared	
  to	
  best	
 ...
Antennas	
   Central	
  Signal	
  
Processing	
  (CSP)	
  
	
  
Transfer	
  antennas	
  to	
  CSP	
  
2020:	
  20,000	
  P...
Revolu8onary	
  sensor	
  data	
  processing	
  
Data	
  
Massive	
  volumes	
  of	
  input	
  
	
  deliver	
  to	
  proce...
Upcoming SlideShare
Loading in …5
×

Processing 1 ExaByte per Day for the SKA Radio Telescope

624 views

Published on

In this deck the Disruptive Technologies Panel at the HPC User Forum, Peter Braam from Cambridge University presents: Processing 1 EB per Day for the SKA Radio Telescope.

"The Square Kilometre Array is an international effort to investigate and develop technologies which will enable us to build an enormous radio astronomy telescope with a million square meters of collecting area."

Watch the video presentation: http://wp.me/p3RLHQ-exw

Sign up for our insideHPC Newsletter: http://insidehpc.com/newsletter

Published in: Technology
  • Be the first to comment

Processing 1 ExaByte per Day for the SKA Radio Telescope

  1. 1. The  Square  Kilometre  Array  (SKA)   Next  Genera8on  radio  telescope  –  compared  to  best   current  instruments  it  offers      …100  8mes  more  sensi8vity,  ~  106  8mes  faster  imaging   100,000’s  of  antenna’s     Many  science  projects  relate  to  gravity     World  wide  science  –  cf.  CERN   Cambridge  University  leads  compu8ng   2015-­‐09   Peter  Braam   1  
  2. 2. Antennas   Central  Signal   Processing  (CSP)     Transfer  antennas  to  CSP   2020:  20,000  PBytes/day   2028:  200,000  PBytes/day     Over  10’s  to  1000’s  kms     Imaging  (SDP)   HPC  problem     2020:  100  PB/day,  300PF   2028:  10,000  PB/day,  30  EF     EB  archive     10’s  to  1000’s  kms   Data  Processing  Pipeline   2015-­‐04   2   World  wide   science     AnalyHcs   Scheduling     Data  replicaHon  
  3. 3. Revolu8onary  sensor  data  processing   Data   Massive  volumes  of  input    deliver  to  processors    cannot  keep  it   Some  data  is  not  precious!     Exa-­‐byte  archive  output  data     World  wide  disseminaHon  and   scheduling  of  data  analyHcs   through  cloud   Soware   Telescope  age  ~50  years     SoTware  must  adapt  to   1.  new  hardware   2.  new  algorithms   Approach:  advanced   programming  –  data  flow,   automaHc  opHmizaHon,   domain  specific  languages      

×