Research Article

LSTM-Based Traffic Load Balancing and Resource Allocation for an Edge System

Table 1

Notation: list of symbols used in the analysis.

SymbolDescription

Input parameters
Maximum number of containers hosted by the MEC server, indexed by
Number of BS, indexed by
Aggregate computational workload
Time slot duration
Amount of aggregate workload at the input buffer
Amount of aggregate workload at the output buffer
Workload buffers maximum capacity
Maximum processing rate for container
Static energy consumed by container in the idle state
Maximum energy consumed by container at maximum processing rate
Per-container reconfiguration cost caused by a unit-size frequency switching
Workload fraction to be computed by the -th container
Maximum computation load per container
Maximum per-slot and per-container allowed processing time
Energy drained by the NIC when powered, with no data transfer
Maximum number of multiple fast tunable lasers
Maximum energy buffer capacity
Upper and lower energy buffer thresholds

Variables
Total BS energy consumption at time slot
Server’s energy consumption at time slot
Energy drained due to the active containers, w.r.t CPU utilization, at time slot
Energy drained due to container switching the processing rates at time slot
Energy induced by the TOE at
Energy drained due to the virtual-link communication cost at time slot
Energy drained due to the number of active transmission drivers at time slot
Total energy cost incurred by the content caching process
Number of containers to be active in time slot
Instantaneous processing rate
Load-dependent factor
-th virtual link communication rate at slot
BS switching status indicator at
Maximum energy drained by the TOE at
The expected processing time
Number of active transmission drivers at
Energy buffer level in slot
Harvested energy profile in slot
Purchased grid energy in slot