Coordinators:

Alberto Pace, CERN
Bernd Panzer, CERN

This series of lectures addresses the broad domain of data storage and management technologies. It starts by setting the scene and surveying the various data storage media, comparing their major features including performance and cost, as well as discussing their interconnections.

Then, the series describes possible data storage architectures and the associated software solutions. Focusing on Large Data Centres, it addresses the issues of heating and power consumption.

This is followed by a description of storage models and addresses data management issues and their supporting techniques and tools.

Finally, the series  focuses on reliability and performance of modern Data storage systems.

In the course of the series, elements of computer security an authentication that are relevant to data management are also presented.

The series of lectures is complemented by 5 hours of practical exercises on aspects such as Performance Tuning and Peer-to-Peer storage.

Series

Type

Lecture

Description

Lecturer

Data Technologies

Lectures

Lecture 1

Setting the scene: Storage technologies

This lecture is an Introduction to storage technologies. It presents the different types of storage media, discusses their respective performance and provides elements of cost analysis.
Then, the lecture addresses the issue of connectivity between storage elements.

Lecture 2

 Architectures of storage systems

The second lecture describes the possible Architectures for Data Storage systems and the corresponding software solutions.
The lecture focuses on  Large Data Centers, and addresses the central issues of heating and power constraints.

Lecture 3

Computer Security and autentication

Elements of computer security an authentication that are relevant to data management are presented.

Alberto Pace
 

Lecture 4

Models and management

The lecture presents the various Storage Models, and the supporting management  techniques including Name Servers, Interfaces for Data Management (SRM) and Replication tools.
Then , the lecture describes the Data Cache and Garbage Collection policies. It finally addresses Authentication, Authorization, Accounting mechanisms used in data storage systems

Alberto Pace
 

Lecture 5

Performance

The lecture discusses the aspects of reliability and performance of data storage systems.

Il also addresses Peer to Peer Storage.

Alberto Pace

Exercises

Exercise 1 Exercise 2 Exercise 3 Exercise 4 Exercise 5

The first part of hands-on exercises aims to improve understanding of basic parameters in IO systems:

  - network and media latency

  - access patterns

  - OS caching

 - bottlenecks and optimization strategies for local and remote data access.

Few essential Linux tools will be introduced to monitor and measure IO performance avoiding bias introduced by OS caching. Students will experience and measure the impact of latency and access patterns on IO performance.

The second part covers the concept of parallelism and redundancy in storage system. With Measurements we will demonstrate the benefits of parallelism in LAN and WAN environments. The exercises conclude with the implementation and performance tuning of a RAID verification algorithm.

Andreas Peters
Alberto Pace
Bernd Panzer

Prerequisite Knowledge

Desirable prerequisite

and

References to further information

Ability to develop simple programs, basic understanding of networking technologies.