Algorithm 2 Suggest optimal pod placement using coupling and dependency

Require:  DataFrame $counts\_df$, DataFrame $dep\_df$, List of Nodes $nodes$, Mean Threshold $mean\_threshold$, Weight X, Weight Y Ensure:  Optimal Pod Placement $placement$ 1: $norm\_counts\_df\leftarrow counts\_df/counts\_df.values.max\left(\right)$ 2: $norm\_dep\_df\leftarrow dep\_df/dep\_df.values.max\left(\right)$ 3: $combined\_df\leftarrow X\times norm\_counts\_df+Y\times norm\_dep\_df$ 4: $sorted\_pods\leftarrow$ Sort pods by total communication and coupling count in descending order 5: for each $pod1$ in $combined\_df.index$ do 6:     for each $pod2$ in $combined\_df.columns$ do 7:       if $pod1\ne pod2$ and $combined\_df.at[pod1,pod2]>mean\_threshold$ then 8:          Append $(pod1,pod2,combined\_df.at[pod1,pod2\left]\right)$ to $pod\_dependencies$ 9:       end if 10:   end for 11:end for 12:Sort $pod\_coupling$ by combined count in descending order 13:for each $i,pod$ in $enumerate(sorted\_pods)$ do 14:   $node\leftarrow nodes[i\%len(nodes\left)\right]$ 15:   Append $pod$ to $placement\left[node\right]$ 16:end for 17:for each $(pod1,pod2,count)$ in $pod\_coupling$ do 18:   $node1\leftarrow$ Node containing $pod1$ in $placement$ 19:   $node2\leftarrow$ Node containing $pod2$ in $placement$ 20:   if $node1\ne node2$ then 21:     if $\mathrm{len}\left(placement\right[node1\left]\right)<\mathrm{len}\left(placement\right[node2\left]\right)$ then 22:        Move $pod2$ to $node1$ 23:        Remove $pod2$ from $node2$ 24:     else 25:        Move $pod1$ to $node2$ 26:        Remove $pod1$ from $node1$ 27:     end if 28:   end if 29:end for 30:return  $placement$