Welcome to the Scalla/XRootD webpage at CERN.

This software framework is a fully generic suite for fast, low latency and scalable data access, which can serve natively any kind of data, organized as a hierarchical filesystem-like namespace, based on the concept of directory.
As a general rule, particular emphasis has been put in the quality of the core software parts.

The XRootD project started at the beginning of 2003 as a collaboration between SLAC and INFN, in the context of the BaBar experiment.
Not much time later it was given important contributions coming from CERN, and was bundled in ROOT as a generic platform for distributed data access, very well suited for the kind of usage ROOT does.

Scalla means Structured Cluster Architecture for Low Latency Access.
This is the relatively new name given to the whole suite of tools which are part of the (formerly called) XRootD distribution.
The main reason for this name was that "xrootd" is just the name of the data access daemon. Although fundamental, it is just a part of the whole suite.

This website is meant to be a place where to gather various useful information about the Scalla/XRootD suite, download the latest source code and get in touch.

The project was started with some general guidelines, which up to now have been followed by the developers and the various contributors. The aim of the project is to make possible the creation of data repositories with no reasonable size limit, with high data access performance and linear scaling capabilities
(i.e. multiply by 100 the size of the setup and you will get 100 times the performance).
At a glance, here are the very basic items.

High speed access to experimental data
No compromises on access speed and scalability
Write once read many times processing mode
Small block sparse random access
High transaction rate with rapid request dispersement (fast opens, low latency)
Fault tolerance (if servers go, the clients do not die)
Fault tolerance (able to manage in realtime distributed replicas)

Generic Mass Storage System Interface (HPSS, RALMSS, Castor, etc)
Full POSIX access
Server clustering for scalability
Up to 262000 servers per cluster
WAN metaclusters
High WAN data access efficiency (exploit the throughput of modern WANs for direct data access, and for copying files as well)

High efficiency data server (low CPU/byte overhead, small memory footprint)
Configuration requirements scale linearly with site complexity
No 3rd party software needed (avoids messy dependencies)
Low administration cost
Non-Assisted fault-tolerance
Self-organizing servers remove need for configuration changes in big clusters
No database requirements (no backup/recovery issues, high performance)

Our commitment is towards the highest code quality we can provide, hence, one of the developer's stronger points is that
making our software work without tweaks under a broad range of UNIX-like platforms makes up a better software.
For now, the supported platforms are in principle:
- Any reasonably recent Linux distribution (Rule of thumb: if you have gcc <3.x then for sure your distribution is very old)
- Solaris
- Maybe others, please report your experiences

The project would not have been so successful without the effort and the enthusiasm of the various people who worked full-time or contributed to it in various ways.
Here is the list of the main ones we wish to acknowledge. If you fell unsatisfied because your name is not in this list, please contact us and you will be added.

Andy Hanushevsky, Fabrizio Furano, Alvise Dorigo,
Fons Rademakers, Gerri Ganis, Rene Brun, Bertrand Bellenot
Andreas Peters, Derek Feichtinger, Guenter Kickinger
Pavel Jackl, Jerome Lauret, Gregory Sharp
Jacek Becla, Tofigh Azemoon, Wilko Kroeger, Bill Weeks
Peter Elmer
Operational collaborators
BNL, CNAF, FZK, INFN, IN2P3, RAL, SLAC, CERN