Quad Charts

1
Quad Charts

• David L. Mills
• University of Delaware
• http//www.eecis.udel.edu/mills
• mills_at_udel.edu

2
Scalable, High Speed Time Synchronization (D012)
New ideas Dynamic peer discovery using engineered
multicast schemes Automatic, authenticated peer
selection for ubiquitous distributed
protocols Multiple overlapping hierarchical
subtree organization for redundancy and
diversity Maximize service performance relative
to crafted metric and constraints
S1
S2
S1
S1
S2
S1
S2


S2
S2
S2
Clients
Clients
S1
S2
S1
S2
S1
S2



S2
S2
S2
Clients
to buddy (S2)
Impact Minimize dependence on prior configuration
data Avoid intricate case-by-case analysis of
failure/fallback/recovery scenarios Provide
automatic reconfiguration in case of network
reconfiguration or failure
Schedule (second year) Test and refine NTP
autokey authentication scheme Test and refine NTP
autoconfigure scheme Complete NTPv4 beta test
and deploy
University of Delaware David L. Mills
3
Survivable Real-Time Network Services (G409/J175)
New ideas Dynamic peer discovery using
intelligent multicast algorithms Self-organizing,
multiple server topology for redundancy and
diversity Automatic repair and restoral using
adaptive, constrained-metric heuristics Automatic
stratum assignment
Redundant Primary Servers
S1
S1
S1
S1
S2
S2
S2
S2
S2
S2
Diversity, repairable links
S3
S3
S3
Autonomous topology organization
S4
Constrained-metric heuristic
Automatic stratum assignment with dynamic
security associations
Impact Minimize dependence on engineered network
configuration data Avoid intricate case-by-case
analysis of failure/fallback/recovery
scenarios Provide automatic reconfiguration in
case of network reconfiguration or failure
Schedule (third year) Implement and test
asynchronous certificate retrieval for dynamic
re-keying Implement expanding-ring, hop-limited,
manycast algorithms Enhance and add new protocols
to the network simulator for very large networks
University of Delaware David L. Mills
4
Survivable Real-Time Network Services (J175)
New ideas Dynamic peer discovery using
intelligent multicast algorithms Self-organizing,
multiple server topology for redundancy and
diversity Automatic repair and restoral using
adaptive, constrained-metric heuristics Self-organ
izing security infrastructure
Redundant Primary Servers
S1
S1
S1
S1
S2
S2
S2
S2
S2
S2
Diversity, repairable links
S3
S3
S3
Autonomous topology organization
S4
Constrained-metric heuristic
Automatic stratum assignment with dynamic
security associations
Impact Minimize dependence on engineered network
configuration data Avoid intricate case-by-case
analysis of failure/fallback/recovery
scenarios Provide automatic reconfiguration in
case of network reconfiguration or failure
Schedule (fourth year) Refine and test
asynchronous security protocol for timing
networks Refine expanding-ring, hop-limited,
manycast algorithms Develop fractal network
models using our simulator for very large networks
University of Delaware David L. Mills