Communication in the Infosphere: Grid Mobility
If a grid is a distributed structure in which each node “weighs” slightly, but the distribution of tasks to millions of such nodes leads to a quick solution, then why not tie the grid to mobile devices?
Then the problem arises - how to adjust the channel from point A to point B, if both points move continuously? In cellular communication there is a fixed base station, which is connected to a fixed switch. This is how the search hierarchy is formed: on the switch, the table is continuously updated to which base station one subscriber is assigned to.
Simplistically, I know, but such a model will do for us.
Now imagine that there is no fixed switch and no fixed base stations. A switch is a task flow solved in a grid environment; the base station is completely absent as a class.
')
And how to solve the problem of positioning?
This algorithm is drawn:
1. Each point registers in the internal table all the other points that are in the zone of its access (direct connection).
Radius, by the way. should not be great - otherwise in crowded places you can fuck such a sign ... though ... If one person at the minimum takes 0.3 square meters. meters, then for a point with a reception radius of 1 km. we have 10.5 million records - maximum. This is if they stand like a herring in a barrel. Then, with an identifier size of 4 bytes, we have a database of 40 MB. A lot, of course ...
But it is when people are pushed to the maximum. And so the average density of about 1 person per 10-15 square meters. m., which gives us a 30-fold decrease in identifiers. This is 1 MB.
Already tolerated. If we consider the current breeding activity, which can be incorporated in the grid, it is more than tolerable. And if you remember that such a radius is not necessary, and may be enough, 200 m. - normally, in general.
2. If there is an addressee in the table, the message is given to him. In the absence of - a small number (N) of the most distant users, selected in the table, is selected and the message is transmitted to them.
3. The procedure is repeated until the recipient with the correct ID receives the message.
Here is such a thing is drawn. An interesting principle, but technically releasing it is a terrible confusion ...
Then the problem arises - how to adjust the channel from point A to point B, if both points move continuously? In cellular communication there is a fixed base station, which is connected to a fixed switch. This is how the search hierarchy is formed: on the switch, the table is continuously updated to which base station one subscriber is assigned to.
Simplistically, I know, but such a model will do for us.
Now imagine that there is no fixed switch and no fixed base stations. A switch is a task flow solved in a grid environment; the base station is completely absent as a class.
')
And how to solve the problem of positioning?
This algorithm is drawn:
1. Each point registers in the internal table all the other points that are in the zone of its access (direct connection).
Radius, by the way. should not be great - otherwise in crowded places you can fuck such a sign ... though ... If one person at the minimum takes 0.3 square meters. meters, then for a point with a reception radius of 1 km. we have 10.5 million records - maximum. This is if they stand like a herring in a barrel. Then, with an identifier size of 4 bytes, we have a database of 40 MB. A lot, of course ...
But it is when people are pushed to the maximum. And so the average density of about 1 person per 10-15 square meters. m., which gives us a 30-fold decrease in identifiers. This is 1 MB.
Already tolerated. If we consider the current breeding activity, which can be incorporated in the grid, it is more than tolerable. And if you remember that such a radius is not necessary, and may be enough, 200 m. - normally, in general.
2. If there is an addressee in the table, the message is given to him. In the absence of - a small number (N) of the most distant users, selected in the table, is selected and the message is transmitted to them.
3. The procedure is repeated until the recipient with the correct ID receives the message.
Here is such a thing is drawn. An interesting principle, but technically releasing it is a terrible confusion ...
Source: https://habr.com/ru/post/904/
All Articles