Risk Intelligence: Making Profit from Uncertainty in Data Processing System

Algorithm 2

Algorithm for risk-aware assignment.

Input: available Nodes 𝑁=𝑛₁,𝑛₂,𝑛₃,…;

(1) Finding the straggler task 𝑠 by prediction;

(2)  for  each available node 𝑛_𝑖 in 𝑁  do

(3)  Make prediction 𝑇=𝑡_min,…,𝑡_𝑖,…,𝑡_max;

(4)   Compute profit 𝑃_𝑖←𝐸(𝑇);

(5)   Record the maximal 𝑃_𝑖 and node 𝑛;

(6)  end for

(7)  if  remaining time 𝑡_left> the worst reexecution time 𝑡_max  then

(8)   kill task 𝑡𝑎𝑠𝑘⋅𝑘𝑖𝑙𝑙(𝑠);

(9)  end if

(10) assign backup task 𝑇𝑎𝑠𝑘𝑎𝑠𝑠𝑖𝑔𝑛(𝑠,𝑛);