In general BeeGFS is a network file system. It allows clients to communicate with storage servers
via network (anything with TCP/IP on it – or InfiniBand with native VERBS support). A BeeGFS
solution could therefore be called a Network Attached Storage (NAS).
Furthermore, BeeGFS is a parallel file system – by adding more servers, the capacity and
performance of them is aggregated in a single namespace. That way the filesystem performance
and capacity can be scaled to the level which is required for the specific application. Furthermore
BeeGFS is splitting MetaData from ObjectData. The ObjectData is the data users want to store –
whereas the MetaData is the “data about data”, such as access rights and file size – but most
important in the MetaData is the information, on which of the numerous storage servers the
content of a file can be found. The moment a client has got the MetaData for a specific file or
directory, it can talk directly to the ObjectDataServers to retrieve the information – there is no
further involvement of the MetaDataServer (unless MetaData has to be changed).
The number of ObjectStorageServers as well as the number of MetaDataServers can be scaled.
Therefor all aspects of performance requirements can be satisfied by scaling to the appropriate
number of servers.
BeeGFS is targeted at everyone, who has a need for large and/or fast storage. Traditional fields
are HighPerformance- and HighThroughputComputing as well as for storage of (large and
growing) research data. The concept of scalability additionally allows users with a fast (but
perhaps irregular) growth to adapt easily to the situation they are facing over time.
An introduction to BeeGFS 5
A very important differentiator for BeeGFS is the ease-of-use. This was one main design goal
from the beginning. The philosophy behind BeeGFS is to make the hurdles as low as possible
and allow the use of a parallel file system to as many people as possible for their work. We will
explain how this is achieved in the following paragraphs.
The BeeGFS file system comes without licence fee: It is a “free to use” product for end users –
so whoever wants to try it for his own use, can download it from www.beegfs.com and use it.
The client is published under GPL, and the server is covered by the Fraunhofer EULA.
There is professional support available through ThinkParQ to enable customers to use the
product in productive environments with defined response times for support. Also professional
support allows system integrators to build solutions using BeeGFS for their customers.
BeeGFS is a Linux based file system – so all parts (clients, servers) run under Linux operating
systems. There might be native support for other architectures in the future, but there hasn’t
been an implementation so far. Since the client is GPL, everyone should feel encouraged to port
it to a different operating system, if he sees any need for it.