Predicting and deleting idle remote sessions in a distributed file system
AssigneeEMC IP Holding Company LLC
InventorsTejas Anil Vaykole, Pratik Rupala, Kaushik Gupta
Life of this patent
Estimating the term…
Abstract
Techniques are provided for predicting and deleting idle remote sessions in a distributed file system. In an example, a server that serves a plurality of server message block (SMB) sessions analyzes with linear regression techniques those sessions to determine which sessions will become idle. In some examples, the server can then proactively close a SMB session that it predicts that an administrator would choose to close after a given idle period of time, or recommend to an administrator that the SMB session can be closed.
Claims20
A system, comprising:
- a processor;
- and
- a memory that stores executable instructions that, when executed by the processor, facilitate performance of operations, comprising:
- in response to receiving first user action data indicative of listing idle server message block (SMB) sessions that are served by a computing cluster, performing, with a prediction model, a linear regression on information about the SMB sessions to determine idle SMB sessions of the SMB sessions;
- providing identified SMB sessions in a user interface in a precedence order based on the linear regression, wherein the precedence order indicates a recommended order to close the identified SMB sessions;
- in response to receiving second user action data from the user interface indicative of taking an action with respect to a SMB session of the identified SMB sessions, updating, with the prediction model, the linear regression to produce an updated linear regression;
- and
- storing the updated linear regression for the prediction model in the memory.
A method, comprising:
- performing, by a system comprising a processor, a linear regression on information about a first server message block (SMB) session to determine that the first SMB session is idle;
- providing a list of SMB sessions that comprises the first SMB session in a user interface, based on the linear regression;
- in response to receiving user action data from the user interface indicative of taking an action with respect to the list of SMB sessions, updating the linear regression to produce an updated linear regression;
- and
- storing, by the system, the updated linear regression in a computer memory.
A non-transitory computer-readable medium comprising instructions that, in response to execution, cause a system comprising a processor to perform operations, comprising:
- performing a linear regression on information about a group of server message block (SMB) sessions to determine that a first SMB session of a group of SMB sessions is idle;
- providing information about the group of SMB sessions in a user interface, based on the linear regression;
- in response to receiving action data at the user interface indicative of an action to perform with respect to a first SMB session of the group of SMB sessions, updating the linear regression to produce an updated linear regression;
- and
- storing the updated linear regression in a memory.
Dependencies are inferred from the claim text.