Server clustering is essential in today's business environment, in order to provide the high availability/scalability of services required to support 24x7x365 production operations. This high availability/scalability requirement encompasses network operating systems, application services and LAN/WAN network resilience. Today's computing architectures must be designed within an infrastructure combining High Availability, Scalability, Manageability, Flexibility and Cost Effectiveness.

eVolution provides technical consultancy resources to effectively implement Novell or Microsoft Cluster Services, Citrix Server Farms or CheckPoint server load balancing to provide a high availability computing architecture within your organisation.

What is a cluster? A cluster is two or more interconnected servers [nodes] that create a solution to provide higher availability, higher scalability or both. Clustering servers for high availability is seen if one node fails, another node in the cluster assumes the workload of the failed node, and users see no interruption of access. Clustering servers for high scalability includes increased application performance and support for a greater number of concurrent users. Clustering can be implemented at different levels of the system, including operating systems, middleware and application level. The more layers that incorporate clustering technology, the more reliable, scalable and manageable the cluster.

Cluster technology requires the necessary server hardware [nodes], shared storage devices, interconnects and cluster services software/systems management and cluster resources [applications and services].

Novell Cluster Services

Since the release of Netware 5.0, Novell has provided Novell Cluster Services [NCS] with two node clusters included with the base Network Operating System and additional licenses required for larger server clusters.

NCS is a multi-node clustering technology that is integrated with Novell eDirectory. NCS provides high availability and manageability of critical network resources including data [server volumes], applications and network services. NCS supports failover, failback and migration [static load balancing] of individually managed cluster resources.

NCS permits the configuration up to 32 Netware Servers into a high availability Active-Active/Active-Passive cluster, where resources can be dynamically switched or moved to any server in the cluster. Cluster resources can be configured to automatically switch to another node in the event of server failure. Resources can also be switched manually if necessary to troubleshoot hardware or balance server workload.

NCS does not require any proprietary hardware as it can be installed on off-the-shelf Intel architecture servers and generic Storage Area Networks [SANs].

Third-party SAN solutions tested and approved for Netware and Novell Cluster Services include HP StorageWorks, IBM TotalStorage and Dell PowerVault.

Microsoft Cluster Services

Windows Advanced/Datacentre server editions support three basic types of clustering technology namely, Server Clusters, Network Load Balanced  [NLB] and Component Load Balancing [CLB] Clusters.

Server Clusters, or failover clusters, are based upon two or more nodes sharing SCSI or Fibre Channel physical storage. One node can manage all disks or each node can manage one or more disks. Upon failure of one node, another node will take control of the disk resources, providing only a momentary lapse in operation. Server Clusters can be configured as either an Active-Passive or Active-Active Cluster. Server Clusters are generally deployed to eliminate single points of failure for read/write data from users or applications.

Network Load Balanced [NLB] Clusters differ from Server Clusters in that they do not share resources. Each NLB node stores and maintains it's own data. Consistency and data integrity between NLB nodes is essential, as all nodes should be identical. NLB clusters are typically used to provide high availability and load balancing for read-only data.

Component Load Balanced [CLB] Clusters allow applications built using COM+ components to operate across multiple servers, similar to NLB Clusters. However, unlike NLB Clusters, CLB Clusters load balance applications and not servers. The CLB cluster service uses a round-robin technique to distribute application requests to COM+ servers in the cluster.

In addition to the cluster service technology described above, there also exists other methods of providing access to Network Load Balanced Servers [identical mirrored servers], such as Citrix Load Balanced Server Farms and CheckPoint Load Balancing.

Citrix Load Balanced Server Farms

Citrix load-balanced server farms were developed to provide thin client access to Network Load Balanced [NLB] servers. Citrix MetaFrame provides enterprise level centralised computing by providing the latest Windows-based applications  to any client device over any connection. Servers within a server farm can be taken offline without causing any application downtime or loss of productivity for users. Additional servers can also be added to the server farm whenever necessary to increase the number of concurrent users supported. Citrix Installation Management Services is used to install applications from a central location onto multiple load-balanced Citrix Servers.

CheckPoint Load Balancing

CheckPoint VPN-1/Firewall-1 can also provide web server load balancing at the firewall by utilising a logical server to determine which requests go to which physical server in a server farm - a group of mirrored servers.  CheckPoint load balancing is also known as ConnectControl, which is the name of the license required to activate this facility. Load balancing requires two components namely, the load balancing  methods [which includes the lhttp daemon] and the five algorithms. The lhttp daemon is used with the HTTP logical server to redirect HTTP connections to a physical server in the server farm. The five algorithms used by the logical server to determine which physical server gets the load-balanced connection are Server Load, Round Trip, Round Robin, Random or Domain.

Many server vendors provide a dedicated platform-specific [Microsoft, Novell, Linux, Unix] clustering solution, providing all the available rack-mounted server hardware and shared storage devices required to successful build and deploy cluster services within your organisation. However, even for an entry-level or mid-level cluster installation, with the increasing popularity, efficiency, reliance and reducing TCO of Storage Area Networks [SANs], it would be prudent to consider a vendor-based clustering solution that would scale to a flexible shared storage SAN in the future. Market leaders providing such a comprehensive clustering solution based upon either SCSI or Fibre Channel shared storage are IBM [Clustering and TotalStorage] and HP [Clustering and StorageWorks]. For further information, please access the website links below.

Following the design and implementation of your chosen server clustering model, eVolution can provide technically qualified engineers to maintain, manage, configure and troubleshoot  Microsoft or Novell Cluster Services, Citrix Load Balanced Server Farms, CheckPoint Load Balancing services or cluster hardware, including server nodes, SCSI or Fibre Channel shared storage devices, cluster controllers, HBAs and implement/support shared storage cluster backup to maintain optimum flexibility and efficiency of your organisation's server clustering infrastructure.

eVolution has the commercial experience and technical competence invested within it's System Engineers to successfully manage your strategic IT integration project or programme.

To assure you of our dedicated management and support of your project,  unlike other IT computer consultancies, eVolution guarantee only to contract to a single client organisation at a time, consequently their is never a conflict of interest between competing clients for our valuable technical and managerial expertise. A true One2One business relationship.

Please visit www.novell.com for further information on Novell Cluster Services.

Please visit www.microsoft.com for further information on Microsoft Cluster Services.

Please visit www.citrix.com for details of Citrix Load Balanced Server Farms.

Please visit www.checkpoint.com for details of CheckPoint Load Balancing Services.

Please visit www.ibm.com for information on IBM xSeries Clustering products and services.

Please visit www.hp.com for information on HP/Compaq Clustering products and services.

Contact Details :

                                E-mail :             info@evolution-is.com for further details of services, quotations and availability.

                                                         webmaster@evolution-is.com for comments and questions about this Web Site.

                                Telephone :      +44 (0)7810 078141                            Facsimile :    +44 (0)1902 843047

                                Mail :                eVolution Information Systems, Central House, 582-586 Kingsbury Road, Birmingham B24 9ND England.