Laxcus Big Data Management system provides computing power based on diffuse/converge distribution algorithm. The detailed introduction of the algorithm is described in the article "Laxcus: Big Data processing System". This diagram shows the process of generating, sorting, displaying, and storing random numbers in a clustered environment.
The text interpretation of the terminal window in the figure:
<1> "conduct", laxcus system keyword, is the statement description of Diffuse/converge distribution algorithm.
<2> "System_sort", distributed computing middleware naming, names are case-insensitive but require unique in the cluster.
<3> "From, to, collect", the keyword under the conduct command, diffuse/converge the phase of the distribution calculation.
<4> "Sites", conduct keyword, the number of nodes required (node is a logical computer).
<5> "WriteTo", conduct keyword, indicates that the data is written to the file name.
<6> "Begin, end, total, by", the user's custom keyword, the format is "name (data type) = parameter". These keywords are defined by the user and are parsed and processed by the user in their own middleware. The "ORDER by" in the figure is a string type, with arguments and single quotation marks included (compatible with SQL format), and other integer types.
Full Description:
This is a demonstration column that produces individual randomly generated numbers by multiple computers and then combines the coordinates to sort. The middleware computer component named "System_sort", according to the algorithm principle of Diffuse/converge distributing computer, executes the process through terminal action to the middleware of the cluster.
In the from phase, the system starts 6 computers, each of which allocates one-sixth of the total number of units, generating random numbers from 100000 to 9999999. To stage there are 3 computers, to take the number generated from the stage, each computer allocated an average of one-third of the number of total, and the assigned numbers are sorted, sorted in descending mode. The data results are displayed on the terminal and written to a disk file called "/notes/records.bin" (this is a running implementation on a Linux system, if the terminal is running on a Windows system, the corresponding directory structure needs to be changed).
Another: in the actual operation process, the Laxcus Big Data system will check the parameters of the number of middleware nodes required, if not meet the requirements, will automatically reduced to the actual number of available.
Demonstration example of Laxcus Big Data distribution calculation
