AI & ML
N-1 Security constrained Hosting Capacity Assessment for Ultra Large-scale Data Center with Workload Flexibility
arXiv cs.SY0 views
arXiv:2611.24493v1 Announce Type: new Abstract: Transmission network security can substantially alter where large-scale data centers can connect and how much workload flexibility can unlock. This paper develops a hosting-capacity framework that couples computing demand, batch workload scheduling, and N-1 security constraints. A modified IEEE 118-bus case study reveals that two data center sites with identical hosting capacities under normal operation differ by 874.5 MW under the contingencies operation when workload flexibility is enabled. Six-hour batch deferral increases N-1 hosting capacity by up to 24.5\% relative to immediate execution. These gains arise from reducing demand at critical hours while accommodating deferred tasks within subsequent transmission limits and deadlines. At bus 92, a four-hour window captures 90.9\% of the six-hour capacity gain. The findings establish workload flexibility as a location-dependent connection resource whose capacity value requires joint assessment of contingency exposure and deadline constrainted workload recovery, beyond the power reduction available at a single critical hour.
Source:https://arxiv.org/abs/2611.24493PreviousKeeping Shielded Data-Driven Controllers Safe Under System Evolution via Model ConsistencyNextQuantifying the hosting capacity value of workload flexibility for AI data centers: A two-stage electricity-computation co-optimization framework
Related posts

How many small business websites block AI crawlers?
0dev.to WebDev
Our health check said "db ok" for 40 minutes. It was talking to a cache.
0dev.to WebDev
Banco do Brasil: acesso indevido pode ter exposto CPF, e-mail e dados bancários
0dev.to Security

When the Attacker Has No Return Address
0dev.to Security
La nube te mantiene responsable de todos modos; lo local simplemente nunca comet
0dev.to Security
Interrupções em bancos e redes da NSA são ligadas a correções de falhas por IA
0dev.to Security