Transcript Document

Scalability and Development of Space Networks
Vincenzo Liberatore, Ph.D.
Disclaimer: the views expressed here are solely the author’s,
not the presenter’s
Scalability
Definition
• Ability of a system to sustain seamless
operations when certain parameters increase
Dimensions
• Specified across four dimensions
Scalability: Dimensions
Dimension
Definition
Numerical
Increased number of
users, resources, and
services
Geographical
Users and resources
that lie far apart
Administrative
Easy to manage even
if it encompasses
multiple administrative
domains
Functional
Increasingly more
complex functionality
Scalability: Terrestrial Background
Dimension
Definition
Terrestrial concern
Numerical
Increased number of
users, resources, and
services
Exponential increase in
number of users, data
volume, services, resources
Geographical
Users and resources
that lie far apart
Worldwide network
Administrative
Easy to manage even
if it encompasses
multiple administrative
domains
Rapid increase in
administrative domains
(e.g., autonomous systems)
Functional
Increasingly more
complex functionality
Complex distributed
applications
Scalable Terrestrial Networks
Scalability is Primary Concern
• Exponential increases of key parameters
• Quality Assurance
– “If it scales, it must be working” [O’Dell]
Expandable and Reusable Solutions
• Convergence layers
– E.g., Internet Protocol (IP)
– Support multiple link, transport layers
• Middleware
– E.g., Resource Discovery
– Simplifies design, development, and deployment of
complex distributed applications
– Reduces costs
– Improves system quality
Scalable Space Networks
Assumption
• Combined approach more powerful than
each in isolation
• Leverage on high readiness terrestrial
technology
• Only a working hypothesis
Gap Analysis
• Terrestrial assumptions may be
inappropriate for space networks
Gap Analysis
Dimension
Definition
Terrestrial concern
Space approach
Numerical
Increased number of
users, resources, and
services
Exponential increase in
number of users, data
volume, services, resources
Sparser set of assets
Geographical
Users and resources
that lie far apart
Worldwide network
Earth, Moon, Mars and
beyond
Administrative
Easy to manage even
if it encompasses
multiple administrative
domains
Rapid increase in
administrative domains
(e.g., autonomous systems)
Smaller number of
administrative domains
Functional
Increasingly more
complex functionality
Complex distributed
applications
Flexible, sustainable,
affordable, and
autonomous
Highly optimized
Gap Analysis
Example
• Numerical scalability
– Vast numbers of terrestrial assets
– Fewer and sparser space assets
Objective
• Reconcile gaps
Process
Resolve Gaps
• Needs explicit process
Spiral Development
• Cycle steps to resolve gaps
• Cycle steps to evaluate alternatives
• Milestones to resolve gaps
Spiral Development
Reprinted from [Bohem 89]
Hypothetical Example: Development Scalability
Determine Objectives
•
•
•
•
•
Flexibility
Sustainability
Affordability
Autonomy
Others?
Determine Constraints
•
•
•
•
Computation
Power
Delays and errors
Others?
Hypothetical Example: Development Scalability
Determine alternatives
• Terrestrial Networks
– Sensor Networks
– Common architectures, interfaces, substrates (ongoing at NSF NeTS)
• Alternative Approach I
– Highly optimized systems
– Hooks for flexibility
• Alternative Approach II
– Reference model:
 Common architectures, interfaces, and substrates
 Compile into highly optimized implementation
 Analogy: distributed shared memory
• Alternative Approach III
– Anyone?
Hypothetical Example: Development Scalability
Remaining spiral stages
• Evaluate alternatives, identify resolve risk
• Develop, verify next level product
• Plan next phases
Conclusions
Concern with scalability central to terrestrial networks
Reconcile with space objectives
• Identify and resolve gaps
• Process: Theory W spiral