Instance creation in a computing system
A system and method for efficiently creating and managing application instances in distributed computing systems is disclosed. Controls are presented for specifying an application for instantiation, a data file for use with the application, and a destination for results from the application. Application resources and topology may be recommended to the user based on prior application execution, and CPU, GPU, and interconnect parameters such as bandwidth and latency. The controls may enable to user to customize the recommendations prior to automated instantiation based on the user's needs, such as whether the application is to be run in batch mode or interactive mode.
1. A method for creating instances of applications in a distributed computing system, the method comprising:
presenting a user with a first control to specify an application that the user wishes to instantiate;
presenting the user with a second control to specify a data file that the user wishes to use with the application;
presenting the user with a third control to specify a destination for results from the application;
presenting the user with a first set of controls displaying a recommended logical topology for an application instance, wherein the first set of controls permit the user to customize the recommended logical topology; and
presenting the user with a second set of controls displaying a set of available system resources in the distributed computing system, wherein the second set of controls enable the user to select a subset of the available system resources upon which to instantiate; and
instantiating an instance of the user-specified application on the user-selected subset of the available system resources with the recommended logical topology as customized by the user.
2. The method of claim 1 , wherein the instantiating comprises automatically creating and deploying containers for each node in the user-customized recommended logical topology.
3. The method of claim 1 , further comprising providing a role control to the user, wherein the role control permits the user to specify an application role for each instance of the application.
4. The method of claim 1 , wherein the instantiating comprises automatically creating and deploying virtual machines to the user-selected subset of the available system resources.
5. The method of claim 1 , wherein the first set of controls comprise controls for specifying a selected role for each node, wherein the selected role is selected from a set of options presented based on the application specified by the user.
6. The method of claim 1 , wherein the first set of controls comprise controls for specifying:
a number of master nodes; and
a number of worker nodes.
7. The method of claim 6 , wherein the first set of controls comprise controls for specifying a requested connection type between the master nodes and the worker nodes.
8. The method of claim 6 , further comprising displaying a list of selectable node representations corresponding to the user-customized recommended logical topology, wherein, in response to a node representation being selected, a third set of controls is displayed, wherein the third set of controls display a recommended node configuration and permit the user to customize the recommended node configuration.
9. The method of claim 8 , wherein the third set of controls comprise controls for specifying:
a number of CPUs for the selected node; and
a number of GPUs for the selected node.
10. The method of claim 9 , wherein the third set of controls comprise controls for specifying a minimum bandwidth between the GPUs for the selected node.
11. The method of claim 9 , wherein the third set of controls comprise controls for specifying:
an amount of memory for each CPU for the selected node; and
an amount of memory for each GPU for the selected node.
12. The method of claim 1 , further comprising displaying a predicted execution time based on the user-customized recommended logical topology and size of the user-specified data file.
13. The method of claim 1 , further comprising:
presenting a maximum desired execution time control;
predicting an execution time based on the user-customized recommended topology and the size of the user-specified data file; and
modifying the user-customized recommended topology to meet a user-specified maximum desired execution time.
14. The method of claim 1 , further comprising displaying a predicted execution cost for the user-customized recommended logical topology.
15. The method of claim 6 , wherein the first set of controls comprise controls for specifying an interactive mode connection or a batch mode connection for the application.
16. The method of claim 6 , wherein the first set of controls comprise controls for specifying whether the user expects the application to CPU-bound, GPU-bound, memory-bound, network-bound, storage-bound, or don't know.
17. The method of claim 6 , further comprising presenting the user with an option to save the user-customized recommended logical topology for later reuse or cloning.
18. The method of claim 1 , further comprising presenting the user with interconnectivity attributes for the available system resources, wherein the interconnectivity attributes comprise bandwidth, and latency.
19. A system for scheduling computing jobs, the system comprising:
a plurality of interconnected computing resources;
a management node connected to the plurality of interconnected computing resources, wherein the management node comprises:
a configurator and a deployer, wherein the configurator is configured to:
present a user with a first control to specify an application that the user wishes to instantiate;
present the user with a second control to specify a data file that the user wishes to use with the application;
present the user with a third control to specify a destination for results from execution of the application;
present the user with a first set of controls displaying a recommended topology for an application instance based on the user's input on the first, second, and third controls, wherein the first set of controls permit the user to customize the recommended topology; and
present the user with a second set of controls displaying a set of available system resources of the plurality of interconnected computing resources, wherein the second set of controls enable the user to select a subset of the available system resources upon which to instantiate; and
wherein the deployer is configured to instantiate an instance of the user-specified application on the user-selected subset of the available system resources with the recommended topology as customized by the user.