Codes for Distributed Storage
Short Course on
Organized by
Department of Electrical Engineering, IIT Kanpur
and supported by MHRD under GIAN (Global Initiative of Academic Networks)
July 1st - 10th 2017
Program
| July 1st, 2017 | |
| 09:00 - 10:00 am | Introduction to error correcting codes, representation of information, block codes, concept of errors and erasures |
| 10:00 - 11:00 am | Erasure codes |
| 11:00 - 11:30 am | Tea |
| 11:30 am- 12:30 pm | Finite fields and preliminary mathematical concepts |
| July 2nd, 2017 | |
| 09:00 - 10:00 am | Encoding and decoding of block codes, generator matrix representation, parity check matrix, examples |
| 10:00 - 11:00 am | Bounds on codes |
| 11:00 - 11:30 am | Tea |
| 11:30 am - 12:30 pm | Maximal distance separable codes and related bounds |
| July 3rd, 2017 | |
| 09:00 - 10:00 am | Reed Solomon (RS) codes, hardware architectures for RS codes |
| 10:00 - 11:00 am | Decoding of RS codes, introduction to Jerasure library |
| 11:00 - 11:30 am | Tea |
| 11:30 am - 12:30 pm | Regenerative codes and codes with locality |
| July 4th, 2017 | |
| 09:00 - 10:00 am | Fundamental trade-off between storage and bandwidth in regenerative codes |
| 10:00 - 11:00 am | Codes used XORBAS and Microsoft Azure |
| 11:00 - 11:30 am | Tea |
| 11:30 - 12:30 am | New developments in codes for storage |
| July 5th, 2017 | |
| 09:00 - 10:00 am | Codes for security |
| 10:00 - 11:00 am | Information theoretic bounds on secrecy capacity of codes |
| 11:00 - 11:30 am | Tea |
| 11:30 am - 12:30 pm | Computer organization and file system concepts, memory hierarchy, Input/Output (I/O), performance modelling |
| July 6th, 2017 | |
| 09:00 - 10:00 am | RAID architecture, high performance I/O, disk technologies |
| 10:00 - 11:00 am | Erasure codes in RAID architectures. Implementation exercise on RS codes using Jerasure library | 11:00 - 11:30 am | Tea |
| 11:30 am- 12:30 pm | Computer networks, TCP/IP protocols, sockets |
| July 7th, 2017 | |
| 09:00 am - 10:00 am | Understanding bandwidth and latency |
| 10:00 - 11:00 am | Cluster and cloud computing and distributed storage |
| 11:00 - 11:30 am | Tea |
| 11:30 - 12:30 pm | Introduction to big data, dimension and scaling issues, Hadoop system |
| July 8th, 2017 | |
| 09:00 - 10:00 am | Introduction to Hadoop Distributed File System (HDFS) and its architecture |
| 10:00 - 11:00 am | Practical Hadoop deployment (Multi-node installation), establishing your own Hadoop cluster on cloud |
| 11:00 - 11:30 am | Tea |
| 11:30 am- 12:30 pm | MapReduce paradigm and examples |
| July 9th, 2017 | |
| 09:00 - 10:00 am | Implementing example MapReduce programs on Hadoop |
| 10:00 - 11:00 am | Introduction to Hadoop's erasure code module (HDFS-RAID) and its (14,10) RS code, comparison with Google Colossus RS code | 11:00 - 11:30 am | Tea |
| 11:30 - 12:30 am | Experiments to measure the recovery performance of RS codes in HDFS-RAID |
| July 10th, 2017 | |
| 09:00 - 10:00 am | Practical side of local reconstruction codes. Comparison of XORBAS codes and local reconstruction codes in Microsoft Azure |
| 10:00 - 11:00 am | Experiments involving new codes with different network topology. We will apply MXOR and long MDS codes | 11:00 - 11:30 am | Tea |
| 11:30 am - 12:30 pm | HDFS administration, performance and tuning. Advanced topics (Micro batching and streaming) |