Note that it is not always known precisely. This is because, again, the structure
of Internet is defined globally and hierarchically, while at the local level we are
facing the fact that routing information is limited by the size of routing tables.
Above all, principle design of the whole network as asynchronous system, with no
decision core makes impossible to obtain the precise value of paths. Updating
distances for each routing table executed by networking protocol that serves just
this—refreshing periodically values of distances between nodes and distributing
them to the relevant “customers”—routers.
Therefore, one of the first questions is about how we know the whole distance
between source and destination. Considering network as a graph deciding about
handling a package it is worth to have the whole distance as well as awareness
about the size and structure of the whole network.
Further in special section we will consider some elements of graph theory that
might be applied for implementation of desperation control.
T RT : Time frame is predefined time slot within which a package should or must
reach a destination. We denote is as T RT —time constraints (“real time-ness”) for a
chosen process;
Note that during a package journey we can treat T RT not as a given limit or
requirement but as a parameter that reduces its value along the package passing
through network.
Then T RT is a function of current time for package journey: T RT (t)
T cost : Introduced “The current cost of the journey” is how much time we have spent
so far. Clear, the more we waste time in traffic or queuing the less chances is left to
reach the destination timely.
Naturally, when we perform well and pass through the main path having sufficient time reserve (in terms of T RT ) we should not worry much or be desperate.
Thus:
Desperation ¼ P
X n
i¼1
d i ; T RT ;
X n
j¼1
d j ;
!
ð19:1Þ
Here the first sum represents the whole distance between source and destination;
T RT —time constraints (“real time-ness”) for a chosen process;
The second sum presents the current cost of journey for a process j.
19.3.1 How Desperation Grows
As long as journey of the package continues we are facing, amongst all others, two
big questions to answer:
• How desperation grows, and
• How to handle it
270
19 Distributed Systems: Resilience, Desperation
of Internet is defined globally and hierarchically, while at the local level we are
facing the fact that routing information is limited by the size of routing tables.
Above all, principle design of the whole network as asynchronous system, with no
decision core makes impossible to obtain the precise value of paths. Updating
distances for each routing table executed by networking protocol that serves just
this—refreshing periodically values of distances between nodes and distributing
them to the relevant “customers”—routers.
Therefore, one of the first questions is about how we know the whole distance
between source and destination. Considering network as a graph deciding about
handling a package it is worth to have the whole distance as well as awareness
about the size and structure of the whole network.
Further in special section we will consider some elements of graph theory that
might be applied for implementation of desperation control.
T RT : Time frame is predefined time slot within which a package should or must
reach a destination. We denote is as T RT —time constraints (“real time-ness”) for a
chosen process;
Note that during a package journey we can treat T RT not as a given limit or
requirement but as a parameter that reduces its value along the package passing
through network.
Then T RT is a function of current time for package journey: T RT (t)
T cost : Introduced “The current cost of the journey” is how much time we have spent
so far. Clear, the more we waste time in traffic or queuing the less chances is left to
reach the destination timely.
Naturally, when we perform well and pass through the main path having sufficient time reserve (in terms of T RT ) we should not worry much or be desperate.
Thus:
Desperation ¼ P
X n
i¼1
d i ; T RT ;
X n
j¼1
d j ;
!
ð19:1Þ
Here the first sum represents the whole distance between source and destination;
T RT —time constraints (“real time-ness”) for a chosen process;
The second sum presents the current cost of journey for a process j.
19.3.1 How Desperation Grows
As long as journey of the package continues we are facing, amongst all others, two
big questions to answer:
• How desperation grows, and
• How to handle it
270
19 Distributed Systems: Resilience, Desperation