Mirador: An Active Control Plane for Datacenter Storage
نویسندگان
چکیده
This paper describes Mirador, a dynamic placement service implemented as part of an enterprise scale-out storage product. Mirador is able to encode multidimensional placement goals relating to the performance, failure response, and workload adaptation of the storage system. Using approaches from dynamic constraint satisfaction, Mirador migrates both data and client network connections in order to continuously adapt and improve the configuration of the storage system.
منابع مشابه
Control Plane Design and Performance Analysis for Optical Multicast-capable Datacenter Networks Acknowledgements
Control Plane Design and Performance Analysis for Optical Multicast-Capable Datacenter Networks
متن کاملAn Empirical Study of the Control and Data Planes (or Control Plane Determinism is Key for Replay Debugging Datacenter Applications)
Replay debugging systems enable the reproduction and debugging of non-deterministic failures in production application runs. However, no existing replay system is suitable for datacenter applications like Cassandra, Hadoop, and Hypertable. For these large scale, distributed, and data intensive programs, existing methods either incur excessive production overheads or don’t scale to multi-node, t...
متن کاملTIMELY: RTT-based congestion control for the datacenter – Public Review
The context is datacenter congestion control. Traditional TCP transport stacks fare poorly in this environment, which has led to considerable interest in recent years in developing specialized transports that aim to deliver high bandwidth utilization at extremely low, microsecond-level packet latency. This is important for demanding datacenter applications such as cloud storage and near-realtim...
متن کاملFocus Replay Debugging Effort on the Control Plane
Replay debugging systems enable the reproduction and debugging of non-deterministic failures in production application runs. However, no existing replay system is suitable for datacenter applications like Cassandra, Hadoop, and Hypertable. On these large scale, distributed, and data intensive programs, existing replay methods either incur excessive production recording overheads or are unable t...
متن کاملEnd-to-end Performance Isolation Through Virtual Datacenters
The lack of performance isolation in multi-tenant datacenters at appliances like middleboxes and storage servers results in volatile application performance. To insulate tenants, we propose giving them the abstraction of a dedicated virtual datacenter (VDC). VDCs encapsulate end-to-end throughput guarantees—specified in a new metric based on virtual request cost—that hold across distributed app...
متن کامل