Three Things To Think About

1
(No Transcript)
2
(No Transcript)
3
Three Things To Think About

• Moores Law
• Parkinsons Law
• Wirths Law

4
Question Moores Law states that the speed of
technology doubles every?a) 12 monthsb) 18
monthsc) 24 months
5
Moores Law

• the number of transistors that can be
  inexpensively placed on an integrated circuit is
  increasing exponentially, doubling approximately
  every two years.
• Source Electronics, Volume 38, Number 8, April
  19, 1965

6
Moores Law The Fifth Paradigm
Calculations per Second per 1,000
YEAR
7
Moores LawThe trend continues
THE EVOLUTION OF A REVOLUTION
8
Moores Law

• More Hardware Comparisons
• Cost per transistor Size decrease cost
  decrease
• Computing performance per unit cost.
  Performance per unit cost doubles every 24 months
  
• Hard disk storage cost per unit of information
  Kryders Law rate of increase in hard drive
  capacity roughly equals the rate of increase in
  transistor count. Performance per unit cost
  doubles roughly every 24 months
• RAM storage capacity Possibly increasing at the
  same rate as processing power
• Data per optical fiber Butters Law of
  Photonics the cost of transmitting a bit over
  an optical network decreases ½ every nine months

9
Moores Law

• Popular Myths and Misconceptions
• Random Access Memory (RAM) Speeds and hard drive
  seek times are improved Typically improve only
  a few percentage points per year
• Megahertz Myth popular misconception that the
  clock speed of a processor determines its speed.
  
• Exponential processor transistor growth
  corresponds with proportional exponential
  increases in processing power and speed not a
  11 ratio

10
Moores Law

• Popular Myths and Misconceptions
• Multi-core CPUs increased computing power
  single threaded nature of most applications makes
  this untrue
• Exponentially increased hardware performance
  exponentially increased software performance

11
SO WHY ARENT WE HAPPY?...
12
Parkinsons Law

• The coefficient of inefficiency
• Software Version -
• Software will expand to fill the resources
  available to it

13
Wirths Law

• software gets slower,
• faster
• than hardware gets faster
• Source Niklaus Wirth, 1995

14
Wirths Law

• Software Bloat
• The hardware resource constraints faced by
  programmers in the 1970s have reversed
  resources are now perceived as cheap, leading to
  an environment where efficiency of code is not a
  priority
• Causes
• Broad Specifications
• Higher-level programming languages
• Time-to-market

• Rapid Development
• Disregard for older machines
• Feature competition

15
One Example

• Microsoft windows system requirements

16
How these issues affect the Warfighter

• Expeditionary operations consist of many
  challenges
• Work from the edge backwards
• Limited bandwidth
• Minimal footprint
• Small, hand-held devices
• Requires efficient and tailored code

17
How these issues affect the Warfighter

• Requires open, modular, scalable design
• Virtualize where possible
• Isolate the hardware from the software
• Scale smaller, not larger
• Code software modules for minimal use features

18
Possible Solutions

• Computing Minimalism
• Avoiding Creeping Featurism aka Featuritis
• Avoiding Second-System effect
• YAGNI You Aint Gonna Need It (so dont code
  it)

19
Open Questions?

• How to provide incentives for small software?
• Keep long-standing benchmarks to track bloat
• Provide monetary incentives for lean
• Formalize trade-off between features and
  performance

20
Open Questions?

• What are the second order effects?
• Is performance more important than ltinsert
  favorite metric heregt?
• Does data bloat drive bigger software and
  management?
• Does tailoring for performance impair reuse?

21
Open Questions?

• Where will it all end?
• Limits of physics versus human nature

22
Near-Term Programs To Consider

• DIMHRS Marine Corps
• Total Force Administration System
• Marine Corps Enterprise Information Technology
  Services
• Case Management Solutions

23
Bottom Line

• Simplicity is like virtue. Easy to identify
• but almost impossible to find.
• - Richard Bausch