An SKA costperformance estimation tool

1
An SKA cost/performance estimation tool

• Shinji Horiuchi 1, Peter Hall 2, Aaron Cippendale
  2, John Bunton 2, Steven Tingay 1
• 1 Swinburne University of Technology
• 2 ATNF, CSIRO
• 19-22 July 2004, SKA conference, Penticton

2
Abstract

• A tool to estimate cost and performance of
  proposed SKA technologies isbeing developed. We
  take a simplified linear approach to cost and
  performance for a generic SKA system design.
  Major building blocks such asfeeds, receivers,
  short and long-haul signal transport, correlators
  and computing-costs were defined in a standard
  fashion for use across many concepts, including
  simple hybrid (or composite) solutions. A trial
  of a web-based PHP application for the USA -
  Large N Small Dish (LNSD) concept is
  demonstrated. The plan is to extend the tool to
  all otherconcepts, with concept-specific
  parameterization being a relatively
  smallsub-section of the code. The program can be
  easily combined with adatabase managment system,
  such as MySQL, for more complicated andaccurate
  estimates.

3
Outline of current overview

• Coding strategy
• Program demonstration
• Future direction
• Requirement for the program
• Easy to handle, update and expand to perform
  cost estimate as realistic and accurate as
  possible

4
Coding strategy HTML - PHP based

• PHP Hypertext Pre-processor HTML embedded
  scripting language, not CGI, suited for dynamic
  web application
• Provide quick page display time, high security,
  transparency to the end user
• PHP syntax is similar to that of C, C and Java
• Support Object Oriented concept
• Ability to interface to a wide range of database,
  such as SQL applications, widely used to create
  database-enabled Web applications

5
Spreadsheets v.s. database?

• Spreadsheets e.g. Excel, HTMLJavaScript
• Merits Client-side script, handy, easy to
  distribute
• Demerits hard to manage complex system, not easy
  to expand, always require to modify program
  whenever parameters (usually hidden) are changed
• Database Managing System (DBMS) e.g. SQL
  applications such as Oracle, PostgreSQL, MySQL
• Merits flexibility to update and expand, fast
  process

6
Adding up complicated SKA components for cost
estimate

• Analogous to Entity-Relationship modeling for
  designing the database
• There are constraints that must rigidly enforced
  on the cost/performance entry, such as required
  data rate
• - Database Management System must do good job to
  handle those

7
MySQL command line interface and a simple PHP Web
script
8
3. Current status of the program

• Represents current model reasonably well
• Only for LNSD yet to produce detail performance
  and cost breakdown

9
Structure of the SKACost/Peformance Program Code
--- (lnsd,lunerburg,cylinder,preloaded,lar,
kast,aperature) techlnsd
include(main_lnsd.php)
//Web input for lnsd for ( i 0 ii ) //loop for regions
include(global.php)
include(antenna.php)
include(feed.php)
include(rf-electronics.php)
include(short-haul.php)
include(station-electronics.php)
include(long-haul.php)
include(central.php) techluneburg
//Next iteration
for other tech ..... include(report.php
) //shows peformance and cost(s) of each
technology include(graphic.php) //shows
graphic when required ? ---

to be implemented
10
Here are some examples on how it works
http//astronomy.swin.edu.au/shoriuch/full-loop
_040714_2/
11
Web input example
12
Output example
13
4. Future once we are happy with basic concept

• Integration of the database and the cost
  computation interface
• Implementing graphic capability
• Adopting realistic parameters for the cost
  computation - the concept of better than zero
  (BTZ) estimation for all possible cost
  components
• Extending to other technology, enabling
  comparison