• We suggest to consider that priority of a message (that might include several
packets) should be treated as variable. This variable can change both ways:
decrease or grow along the journey of a package through the network,
depending the network and package states.
• Our approach assumes that:
(A) Introduced limit of permitted processing time for each message is known
(B) Combining with overheads caused by state of internal router priority state
message
knowledge about own progress
Thus, our result in one sentence is:
The proposed new schemes and algorithms increase efficiency of package
handling and overall resilience, real-timeness, elasticity and reliability of
network.
This is kind of system level solution. What else can be done here? Well, in [1–7]
we have proposed a Graph Logic Model that enabled to see monitor incoming and
outcoming conditions for every vertex of graph. It is potentially very useful for
routing as well.
GLM unique features that fit for restructuring of routing table are:
• GLM-upgraded routing table can be adjusted, reflecting change of load and
traffic
• Instead of cost of the link GLM introduces “probabilities” of the link (relative
weight of the link in the network), this enables to estimate much faster the
success of the journey
• Logic operator for each incoming and outcoming links should be modifiable.
Thus, when the system is overloaded that becomes visible by refusing to accept
further incoming traffic—(a good example here is a so-called “denial of service”)
GLM logic operator (for incoming link) can be modified from XOR down to NOP,
and when traffic changes, can return back. For this purposes GLM introduces XOR
(to accept all from this particular direction and nobody else) or NOP (nothing from
this direction is welcomed) operators at the input and outputs of a node. At the node
level the condition to process incoming and output flows might change and routing
table outcome operators accordingly can accept all or several types (AND, OR)
outcomes.
These all combined can handle different packages in various types of network
load and conditions by the most efficient way.
This combination creates a kind of framework for network handling, considering
both: types of packages and router conditions.
19.11 Aftermath—Instead of Conclusion
291
packets) should be treated as variable. This variable can change both ways:
decrease or grow along the journey of a package through the network,
depending the network and package states.
• Our approach assumes that:
(A) Introduced limit of permitted processing time for each message is known
(B) Combining with overheads caused by state of internal router priority state
message
knowledge about own progress
Thus, our result in one sentence is:
The proposed new schemes and algorithms increase efficiency of package
handling and overall resilience, real-timeness, elasticity and reliability of
network.
This is kind of system level solution. What else can be done here? Well, in [1–7]
we have proposed a Graph Logic Model that enabled to see monitor incoming and
outcoming conditions for every vertex of graph. It is potentially very useful for
routing as well.
GLM unique features that fit for restructuring of routing table are:
• GLM-upgraded routing table can be adjusted, reflecting change of load and
traffic
• Instead of cost of the link GLM introduces “probabilities” of the link (relative
weight of the link in the network), this enables to estimate much faster the
success of the journey
• Logic operator for each incoming and outcoming links should be modifiable.
Thus, when the system is overloaded that becomes visible by refusing to accept
further incoming traffic—(a good example here is a so-called “denial of service”)
GLM logic operator (for incoming link) can be modified from XOR down to NOP,
and when traffic changes, can return back. For this purposes GLM introduces XOR
(to accept all from this particular direction and nobody else) or NOP (nothing from
this direction is welcomed) operators at the input and outputs of a node. At the node
level the condition to process incoming and output flows might change and routing
table outcome operators accordingly can accept all or several types (AND, OR)
outcomes.
These all combined can handle different packages in various types of network
load and conditions by the most efficient way.
This combination creates a kind of framework for network handling, considering
both: types of packages and router conditions.
19.11 Aftermath—Instead of Conclusion
291