Data latency evaluation

Abstract

The described technology is generally directed towards methods for data latency evaluation. The techniques disclosed herein can provide useful information about when a data consumer can expect to receive data. Methods can create and compare data latency cumulative probability distributions comprising probabilities associated with different latency values, at various different levels of completeness.

What is claimed is: 
     
         1 . A method, comprising:
 collecting, by a system comprising a processor, data latency values associated with multiple configurations of a network, wherein the data latency values are collected for multiple performance benchmarks and wherein the data latency values are associated with a same level of data delivery completeness;   generating, by the system, multiple cumulative probability distributions using the data latency values, wherein the multiple cumulative probability distributions correspond to the multiple performance benchmarks;   comparing, by the system, the multiple cumulative probability distributions; and   selecting, by the system, based on a result of the comparing, a configuration of the multiple configurations of the network in order to increase a performance of data delivery to a consumer via the network.   
     
     
         2 . The method of  claim 1 , wherein the multiple cumulative probability distributions represent probabilities of observed data latency values being less than or equal to distribution latency values. 
     
     
         3 . The method of  claim 1 , wherein the same level of data delivery completeness comprises a same group of multiple different levels of data completeness. 
     
     
         4 . The method of  claim 1 , wherein the same level of data delivery completeness comprises a partial level of data delivery completeness. 
     
     
         5 . The method of  claim 1 , wherein comparing the multiple cumulative probability distributions comprises comparing a first probability of a first cumulative probability distribution and a second probability of a second cumulative probability distribution, and wherein the first probability and the second probability are associated with a selected distribution latency value. 
     
     
         6 . The method of  claim 1 , wherein comparing the multiple cumulative probability distributions comprises determining whether first probabilities of a first cumulative probability distribution comprise greater probabilities of respective latencies being below selected distribution latency values than corresponding second probabilities of a second cumulative probability distribution. 
     
     
         7 . The method of  claim 1 , wherein the same level of data delivery completeness comprises a minimum amount of data to complete a minimum computation using the data. 
     
     
         8 . The method of  claim 1 , wherein the same level of data delivery completeness comprises an amount of data sufficient to complete a full computation using the data. 
     
     
         9 . A system, comprising:
 a processor; and   a memory that stores executable instructions that, when executed by the processor, perform operations, comprising:   generating multiple cumulative probability distributions using multiple data latency values,   wherein the multiple cumulative probability distributions correspond to multiple performance benchmarks,   wherein the multiple cumulative probability distributions are generated using data latency values associated with multiple configurations applicable to a network, and   wherein the data latency values are associated with a same level of data delivery completeness;   comparing the multiple cumulative probability distributions; and   selecting, based on a result of the comparing, a configuration of the multiple configurations of the network in order to modify a performance characteristic associated with data delivery via the network.   
     
     
         10 . The system of  claim 9 , wherein comparing the multiple cumulative probability distributions comprises determining a respective cumulative probability distribution of the multiple cumulative probability distributions that comprises a greater respective probability of an observed data latency value being below a selected target latency value. 
     
     
         11 . The system of  claim 9 , wherein comparing the multiple cumulative probability distributions comprises comparing distribution probabilities at multiple distribution latency values. 
     
     
         12 . The system of  claim 9 , wherein comparing multiple cumulative probability distributions comprises comparing distribution probabilities at all distribution latency values. 
     
     
         13 . The system of  claim 9 , wherein generating the multiple cumulative probability distributions using multiple data latency values comprises collecting data latency values associated with multiple levels of data delivery completeness. 
     
     
         14 . The system of  claim 9 , wherein generating the multiple cumulative probability distributions comprises conducting the multiple performance benchmarks using multiple different equipment configurations. 
     
     
         15 . The system of  claim 9 , wherein the same level of data delivery completeness comprises a partial level of data delivery completeness. 
     
     
         16 . A non-transitory machine-readable medium, comprising executable instructions that, when executed by a processor of a user equipment, perform operations comprising:
 collecting, for a first performance benchmark, first data latency values associated with data delivery via a first configuration of a network to a consumer, resulting in first collected data latency values;   collecting, for a second performance benchmark, second data latency values associated with data delivery via a second configuration of the network to the consumer, resulting in second collected data latency values,   wherein the first collected data latency values and second collected data latency values are associated with a same level of data delivery completeness;   generating a first and a second cumulative probability distribution, respectively, using the first data latency values and the second data latency values, respectively, wherein the first cumulative probability distribution and the second cumulative probability distribution, respectively, correspond to the first performance benchmark and the second performance benchmark, respectively;   comparing the first cumulative probability distribution with the second cumulative probability distribution; and   selecting, based on a result of the comparing, the first configuration of the network or the second configuration of the network.   
     
     
         17 . The non-transitory machine-readable medium of  claim 16 , wherein the same level of data delivery completeness comprises a minimum amount of data to complete a computation using the data. 
     
     
         18 . The non-transitory machine-readable medium of  claim 16 , wherein the same level of data delivery completeness comprises a partial level of data delivery completeness. 
     
     
         19 . The non-transitory machine-readable medium of  claim 16 , wherein comparing the first cumulative probability distribution with the second cumulative probability distribution comprises determining whether the second cumulative probability distribution is stochastically greater than the first cumulative probability distribution. 
     
     
         20 . The non-transitory machine-readable medium of  claim 16 , wherein the operations further comparing the first cumulative probability distribution or the second cumulative probability distribution with a download requirement for data delivery via the network.