Inferring the Network Latency Requirements of Cloud Tenants

   Abstract