. 2021 Feb 3;7:e342. doi: 10.7717/peerj-cs.342

Algorithm 2. Management of fog nodes.

Step 1

Each fog node sends the

S T_{e s t}^{k}

(for node k) to node_dt

node_dt assigns each

S T_{e s t}^{k}

to the ordered L_node list.

foreach req_i do

req_i is sent to node_dt

(Pr_i) = Priority(

d e l a y_{i}^{T}

S T_{e s t}^{T}

)

if Pr_i = H then

Q_{H} (l a s t) \leftarrow r e q_{i}

else if Pr_i = M then

Q_{M} (l a s t) \leftarrow r e q_{i}

else if Pr_i = L then

Q_{L} (l a s t) \leftarrow r e q_{i}

Step 2

/* Task distribution to fog nodes according to

S T_{e s t}^{k}

and

d e l a y_{i}^{T}

foreach req_i in Q_H, Q_M, Q_L do

for L_node(k) do

if $S T_{e s t}^{k} \geq d e l a y_{i}^{T}$ then

n o d e_{k} \leftarrow r e q_{i}

break

if req_i is not assigned to any fog node then

reversedList

\leftarrow r e v e r s e (L_{n o d e} (k))

/* from high to low (processing capacity) */

s u m \leftarrow L_{n o d e} (f i r s t)

for reversedList do

s u m \geq d e l a y_{i}^{T}

then

L_{n o d e} (f i r s t), L_{n o d e} (s e c o n d), . . ., L_{n o d e} (k) \leftarrow r e q_{i}

break

else

s u m \leftarrow s u m + L_{n o d e} (n e x t)

if req_i is not assigned to any fog node group then

if (req_i in Q_H) or (req_i in Q_M) then

reject(req_i)

else if req_i in Q_L then

c l o u d \leftarrow r e q_{i}