Effective incident management is pivotal for the smooth operation of Microsoft cloud services. In order to expedite incident mitigation, service teams gather troubleshooting knowledge into Troubleshooting Guides (TSGs) accessible to On-Call Engineers (OCEs). While automated pipelines are enabled to resolve the most frequent and easy incidents, there still exist complex incidents that require OCEs’ intervention. In addition, TSGs are often unstructured and incomplete, which requires manual interpretation by OCEs, leading to on-call fatigue and decreased productivity, especially among new-hire OCEs. In this work, we propose Nissist which leverages unstructured TSGs and incident mitigation history to provide proactive incident mitigation suggestions, reducing human intervention. Leveraging Large Language Models (LLM), Nissist extracts knowledge from unstructured TSGs and incident mitigation history, forming a comprehensive knowledge base. Its multi-agent system design enhances proficiency in precisely discerning OCE intents, retrieving relevant information, and delivering systematic plans consecutively. Through our user experiments, we demonstrate that Nissist significantly reduce Time to Mitigate (TTM) in incident mitigation, alleviating operational burdens on OCEs and improving service reliability. Our webpage is available at https://aka.ms/nissist.

A connection-oriented congestion control protocol for broadband ISDN is presented. The protocol employs the traffic viewpoint hierarchical design approach and is implemented by a two-layer scheme at the call and cell layers. The call layer handles call admission and provides shortest path connection routing when a call is admitted, while the cell layer allocates switch input buffer space for each traffic type according to each type's cell loss probability grade-of-service requirement. There is a direct interaction between the two layers in processing a connection request. A discrete-time queuing system with geometrically distributed service time and state-dependent Markov modulated Bernoulli process arrivals is established to model heterogeneous networking environments for this protocol. An analytical solution is developed for this queuing system. Numerical results show that the protocol can reduce network congestion as input traffic rates approach switch capacity, and increase network utilization relative to that possible in the absence of congestion control.>

The authors examine an optimal bandwidth allocation scheme for multichannel switching systems to switch trunk groups. This multichannel switching system features a simple architecture, high throughput, high channel utilization, and other properties such as preserving packet sequence, being robust to unbalanced input traffic loads, providing fair access to network transmission resources, and fast packet processing via a pipeline structure. Numerical results show that the multichannel system can achieve a throughput of 91% per channel for a trunk group of 32 channels while also providing very low average cell delay performance in the presence of correlated inputs.>