Adaptive computation and faster computer operation
A datacenter has more computing power than a personal computer. The personal computer sends a request to perform an operation on a data set to the datacenter. The datacenter evaluates various inputs to determine if, despite the datacenter's computing power, the personal computer is likely to complete the operation faster. Based on the determination, the datacenter may perform the operation, send the data set to the personal computer for the personal computer to process, or start a competitive computation. As a result, a user interface can be more responsive. Machine learning processes can be used to improve predictions.
1. A system comprising:
one or more processors; and
a network communication interface;
wherein the system is configured to:
receive, through the network communication interface, a request from a personal computer to perform a computation on a data set;
access information indicating past speeds of performing the computation on other data sets by the system and the personal computer;
based at least in part on a comparison between the past speeds of performing the computation on other data sets by the system and the personal computer, determine to perform the computation on the data set or transmit the data set to the personal computer; and
in response to the determining, initiate performing the computation on the data set or transmitting the data set to the personal computer.
2. The system of claim 1 , wherein the information indicates whether the system completed the computation on a previous data set before the personal computer completed the computation on the previous data set.
3. The system of claim 1 , wherein the system is further configured to:
perform the computation on the data set or transmit the data set to the personal computer further based at least in part on a hardware specification of the personal computer.
4. The system of claim 1 , wherein the system is further configured to:
perform the computation on the data set or transmit the data set to the personal computer further based at least in part on at least two of:
comparing a size of the data set to a threshold size,
a type of the computation being performed,
a type of data being operated on in the computation,
an identity of a user making the request to perform the computation, or
an identity of the personal computer.
5. The system of claim 1 , wherein the system is further configured to:
perform the computation on the data set in response to receiving the request to perform the computation on the data set; and
transmit a result of the computation to the personal computer.
6. The system of claim 5 , wherein the system is further configured to:
transmit the data set to the personal computer; and
determine whether the system performed the computation on the data set before the personal computer locally performed the computation on the transmitted data set.
7. The system of claim 1 , wherein the system is configured to perform the computation on the data set by executing a large scale data processing engine.
8. The system of claim 7 further comprising:
a plurality of storage devices set up with a fault-tolerant configuration; and
wherein the one or more processors are configured to perform the computation tolerant of a failure by at least one processor of the one or more processors.
9. The system of claim 1 , wherein the system has at least one of: more raw processing power, more random access memory (RAM), more storage capacity, or better redundancy than the personal computer.
10. A method performed by a computer system, the method comprising:
receiving, through a network communication interface, a request from a personal computer to perform a computation on a data set;
accessing information indicating past speeds of performing the computation on other data sets by the computer system and the personal computer;
based at least in part on a comparison between the past speeds of performing the computation on other data sets by the computer system and the personal computer, determining to perform the computation on the data set or transmitting the data set to the personal computer; and
in response to the determining, initiating performing the computation on the data set or transmitting the data set to the personal computer.
11. The method of claim 10 , wherein the information indicates whether the computer system completed the computation on a previous data set before the personal computer completed the computation on the previous data set.
12. The method of claim 10 further comprising:
determining a hardware specification of the personal computer; and
wherein performing the computation on the data set or transmitting the data set to the personal computer is further based in part on the hardware specification of the personal computer.
13. The method of claim 10 further comprising:
selecting between performing the computation on the data set or transmitting the data set to the personal computer based in part on at least two of:
determining that a size of the data set is less than a threshold amount,
a type of the computation,
a type of data being operated on in the computation,
an identity of a user making the request to perform the computation, or
an identity of the personal computer.
14. The method of claim 10 further comprising:
performing the computation on the data set using a one or more processors in response to receiving the request to perform the computation on the data set; and
transmitting a result of the computation to the personal computer.
15. The method of claim 14 further comprising:
transmitting the data set to the personal computer; and
receiving a time that the personal computer finishes performing the computation on the transmitted data set.
16. The method of claim 14 , wherein performing the computation further comprises executing the computation with a large scale data processing engine.
17. The method of claim 14 further comprising:
reading the data set from storage devices set up with a fault-tolerant configuration; and
wherein performing the computation further comprises performing the computation using one or more processors that are configured to perform the computation tolerant of a failure by at least one processor of the one or more processors.
18. The method of claim 14 , wherein the request from the personal computer includes one or more keys that references values of the data set.