GATE uses a highly parallelizable GPU-compatible decomposition to solve traffic engineering problems, achieving near-optimal solutions 5-10x faster than state-of-the-art methods on production traces from large cloud WANs with guaranteed convergence.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GATE: GPU-Accelerated Traffic Engineering for the WAN
GATE uses a highly parallelizable GPU-compatible decomposition to solve traffic engineering problems, achieving near-optimal solutions 5-10x faster than state-of-the-art methods on production traces from large cloud WANs with guaranteed convergence.