IP-Tech migrates to the Diskless Shared Root Cluster: Higher Flexibility and scalabiltiy

A flexible Linux cluster is the key to success

The IP-Tech AG is one of the biggist Full Service Provider in Switzerland. The company provides its customers hosting, voice over IP and applikation services. Founded in 2000 it growth fast and supports more than 25.000 domains for web or/and mail services in the meantime. Over the years the companys IT infrastructure had accrued to more than 30 server, each with its own operating system and storage unit. Facing this heterogenous environment all signs at IP-Tech pointed to homogenisation and consolidation.

The decisive factor for a radically IT restructuring of the Full Service Provider was a project of a new acquired customer. “The needs of our customers are also our needs”, explains Donat Grimm, director of IP-Tech. “For a service provider of the digital pre-press we should operate a media data base, on which he depends on for producing glossy brochures. That does not only require the storage of huge data amounts in the terabyte range, but also the content availability of them. With our previous approach we could not get any further, that is why we were looking for an alternative, which should be clearly more effective, flexible and scalable. Furthermore we wanted to reduce the cost. It was clear, that wie needed a powerful cluster solution. Toghether with ATIX we found the ideal solution for us.”

Commencing with the hardware, a totally new platform on base of the Red Hat Enterprise Linux was created. The heart of the new cluster solution, which IP-Tech had elected finally, represents the Red Hat Global File System (GFS). The GFS is an open source file system, which enables various server in a cluster the parallel access to the same physical block of data. ATIX, Red Had Advanced Partner for GFS, clustering and storage, developed on base of it the “com.oonics Diskless Shared Root Cluster”, which consequently thinks the cluster approach out and which provides IP-Tech an extremely flexible and easy scalable system.

By using a rudimental cluster method, IT services run parallel on various server and the user inquiries are distributed by a load dispatcher. For dynamical services it is neccessary to uncouple the application server from the data management and to enable them a corporate access to the storage system. The traditional approach for a comman data storage is a file server. This server often becomes a single point of failure (SpoF) for the complete system.

In order to overcome these problems inherent in the traditional approach, every server in the cluster would have to have direct access to the storage devices and be able both to read them and write to them concurrently. A cluster file system connects application servers to a linked storage network (Storage Area Network - SAN) and allows parallel file system access from all cluster nodes to a central storage system. The diskless shared root cluster developed by ATIX takes this approach to its logical conclusion and avoids having any hard disks at all in the cluster nodes.

IP-Tech uses for its initial configuration eight HP ProLiant BL20p Blades each with two Intel Processors. “Currently we are using also other chip architecutures, but in medium term we woul like to consolidate to Intel Processors. As industry standard they offer clearly the best cost/performance ratio and the biggest freedom by choosing the operating system and the applications”, explains Grimm. The Red Hat Cluster Suite is responsible on the one hand for ensuring that the services of one server are taken over by another if it fails, and on the other for distributing the load between the cluster nodes.

The Red Hat GFS allows parallel file system access by all cluster nodes to a central storage system, which is based on two HP Enterprise Virtual Arrays (EVA) 3000. The GFS pool layer virtualises the storage devices and makes them available so they are non-hardware-dependent. This means that several devices can be combined in a pool using "striping" or linear concatenation. Any changes to the pool configuration are visible to all cluster servers and a volume manager ensures that the file system can be expanded online. Since, in a GFS storage cluster, many servers are accessing the same physical data blocks, there is a mechanism for coordinating the distributed accesses - the so-called "locking service". This guarantees data consistency within the file system.

The new IP-Tech infrastructure

Aussagen von IP-Tech Performance-Steigerung: x10 Stabilitätsverbesserung: erheblich Kosteneinsparung: 40 %

The innovative clou of the Diskless Shared Root Cluster, developed by ATIX is to completely to do away with hard disks in the HP blades and to boot them directly from the storage system. This configuration is very easily scalable: new resources in the form of new server hardware can simply be added on the "plug & play" principle, because the operating system is also installed centrally on the storage system. In addition, it makes maintaining the operating system much easier, because there is only ever one version of the operating system to update.

The complete separation of cluster nodes and central disk storage means that all information about the structure and content of the cluster is consolidated in the central storage system. This means that if one server fails, no information is affected which has to be reinstated. This reduces the restart time for a cluster node to a minimum, because only the server hardware has to be replaced in order to return the system to its normal status. This increases the overall availability of the cluster. The standard tools of the HP storage system are used for backup. For the backup the features of the EVA storage system were used. The cluster can also be remotely monitored and administered by using the ATIX solution com.oonics GrayHead.

“With the Diskless Shared Root Cluster we found more than a optimal system for our customers” resumes Donat Grimm. “With ists high flexibility and scalability the system provides us with the possibility to react fast and cost-effective on every customer requirement. A competitive system can not be found in the proprietary world. That the solution could only be realised with the Red Hat GFS was, besides the cost, casting for the Open Source technologie. Needless to say, we benefit from the attained vendor independence and the increased performance due to the Intel platforms.”