SKADSLOGO

1
DS3The Network and its Output DataPaul
Alexander
2
Overview
Overall aim of this design study Examine the
System Design issues associated with producing
the most cost-effective overall architectural
design for the SKA network
3
The physical network
Data flow, processing and analysis
signal and data transport
power
4
The Structure of DS3

• Six tasks
• T1 Network infrastructure and data
  transmission Manchester
• T2 Data handling control and distributed
  computing ASTRON
• T3 Architecture and the functional
  simulator Cambridge
• T4 Siting and related issues OPAR
• T5 SKA for the user Cambridge
• T6 Scaleable design and implementation ASTRON

5
DS3 and the SKA system design

• Concerned with the overall SKA architecture not
  just the Aperture Array
• Will consider compare and evolve
• SKADS Benchmark specification
• ISPO reference design
• Other emerging system designs

6
Current Status

• All tasks underway
• No DS deliverables or milestones in the first 12
  month period
• Relatively slow start for some tasks
• Planned in the description of work
• Later than anticipated availability of UK funds
• At kickoff meetings identified
• the need for close inter task discussions and
  also with other DS

7
Some of the interactions
DS3-T1 Network
DS3-T6 Scaleup
DS3-T2 Processing
DS4-T2 DSP
DS2-T2 Science
DS3-T3 Simulator
DS4-T6 2-Pad
DS5 Embrace
DS3-T4 Siting
8
Current Status

• All tasks underway
• No DS deliverables or milestones in the first 12
  month period
• Relatively slow start for some tasks
• Planned in the description of work
• Later than anticipated availability of UK funds
• At kickoff meetings identified
• the need for close inter task discussions and
  also with other DS
• cost control and cost constraints central to
  system design

9
Cost modelling in DS3

• In the original description of work there exist
  many work packages which are required to produce
  explicit cost models for particular aspects of
  SKADS technology
• Cost control central to work in DS3, DS4, DS5 and
  DS6
• Aim to coordinate this work and provide a good
  link to the wider project Single identified
  point of contact to bring together the
  SKADS costing work Provide an interface
  between SKADS costing work and the ISPO
  development of a cost model Communicate
  emerging cost output from SKADS to ISPO
  and communicate relevant aspects of ISPO cost
  model to SKADS team

10
Highlights so far

• DS3-T1 (more in a moment) Network infrastructure
  and data transmission
• A first good costing model for the physical
  network
• DS3-T2 (more in a moment) Data handling, control
  and distributed computing
• DS3-T3 (more in a moment) Overall architecture
  and network simulator
• DS3-T5 (planned start end 06) SKA for the user
• DS3-T6
• Need to adopt good practice for how we model and
  go about the design process especially as we move
  into the Engineering Design phase
• Initial review of available methodologies

11
Highlights DS3-T4

• DS3-T4 Siting and related issues
• Original view of the task
• The production of a full plan for siting and
  related issues for the SKA, encompassing
  environmental issues (including signal
  conditioning, impact and adopting the advanced
  technology to arid lands), radio frequency
  monitoring/protection over a long term and
  potential maintenance regimes.

12
Highlights DS3-T4

• DS3-T4 Siting and related issues
• Significant unforeseeable changes since SKADS
  planned in particular delay of site selection
  well beyond late 2006 as originally envisaged
• Some work originally planned in SKADS will now be
  done by International Project (specifically the
  Site Evaluation Working Group and its Task Group
  on Regulatory Issues, and the ISSC )

13
Highlights DS3-T4

• DS3-T4 Siting and related issues
• SKADS participants played central roles in
  helping to define regulatory requirements for
  radio quiet zones at each possible site
• ASTRON engineers have carried out the extensive
  and homogeneous Radio Frequency Interference
  (RFI) monitoring campaign of all four proposed
  SKA sites
• SKADS participants contributed towards ISPO Memo
  73 on Spectrum Protection Criteria for the SKA

14
DS3-T3 Overall Architecture and Network
simulation
15
DS3-T3 Refining the Benchmark Specification

• Aperture Array stations on long baselines
• Benchmark specification has AA providing n lt 1GHz
  collector throughout. Can we define more
  precisely the requirements of an AA station?
• Consider basic parameters station size, filling
  factor and data rate
• Constraint ability to calibrate the station

16
DS3-T3 Refining the Benchmark Specification

• Consider the ionosphere
• typical size of isoplanatic patch at 300MHz is
  2 degrees
• Calibration is relativelystraightforward if each
  stationbeam subtends an anglesmaller than the
  isoplanaticpatch

17
DS3-T3 Refining the Benchmark Specification

• For a larger beam we need a more complex
  ionospheric model which can be solved for if
  reltively short baselines available as in compact
  core
• On long baselines however, sparse station
  coverage
• d 70 m

18
DS3-T3 Refining the Benchmark Specification

• Need 1 source on average per beam/isoplanatic
  patch to calibrate Poisson statistics to give
  P(1) for 99 of time needs N 10 sq deg-1
• At 1.4GHz, use FIRST counts(unresolved or
  model)
• Require DS 15 mJy for 70m AA,70 efficiency
• f gt 20

Integrated FIRST source count
19
DS3-T3 Refining the Benchmark Specification

• Long baseline AA station
• Diameter 70 m
• Filling factor gt 20
• Data rate
• assume 8-bit digitisation for 300 1000 MHz
• Per square degree full bandwidth 24 Gbs-1
• 250 beams (? 250 sq degrees or 60 uniform
  sensitivity) gives 6 1012 bs-1
• Compare to LNSD station giving 1 sq degree at 1.4
  GHz
• f 10 and 25 13m antennae
• Beam form in central processor
• Data rate 6.4 1012 bs-1

20
DS3 Dynamic range

• Sensitivity from large collecting area is only
  sensible if the telescope can also achieve the
  necessary dynamic range given is FOV, beam shape,
  side lobes etc.
• e.g. Tim Cornwell as investigated in detail the
  effects of the telescope quadrupod on the
  producing asymmetries in the beam and hence
  limiting the dynamic range
• This is a big issue which needs to be carefully
  considered for any antenna used in a given band

21
DS3 The Next 18 Months
22
DS3 Next 18 months

• People now in post for DS3-T1 and T3 in UK and
  very significant ramp-up of these tasks is
  underway confidently expect to meet SKADS
  milestones and deliverables on time
• DS3-T2 is now ramping up as planned
• DS3-T4 effectively complete report will be
  written after ISSC complete site selection
• First SKA for the user workshop planned in next
  4 months (DS3-T5)
• Scaleup and design issues work on track as
  planned

23
(No Transcript)
24
Overview

• Overall aim of this design study
• Examine the System Design issues associated with
  producing the most cost-effective overall
  architectural design for the SKA network
• By Network we mean
• The physical network signal and data transport
  power
• The data network data flow, processing and
  analysis
• The output network telescope response warts and
  all
• The user network how the telescope appears to
  the end-user astronomer

25
The next 18 months

• On track to meet milestones and deliverables
• Phase transfer links (T1)
• Frame work for the network simulation / analysis
  (T3)
• Overview report on site selection (T4)
• Report MDA constructs, interfaces and platforms
  (T6)