Zycko Consultant, Sam Samuel (below) tells DCM that SOA is one of the drivers for change in the data centre

On the face of it this should have little impact on the applications as they will still run on a server – whether it is a ‘real’ physical server or a virtual server.  However, the virtualisation capability (which includes the logical deployment and redeployment of data centre resources such as storage, servers and networks) has allowed the advent of Software as a Service – SaaS.

SaaS is a service which allows applications to be run by an outsourced supplier and the applications paid for by usage (this includes time used, processing power consumed, data storage capacity used etc.).  This approach was trialled about a decade ago with ASPs –  Application Service Providers being the next great thing. However, the ASP initiative never took off.  This is because applications had to be run on multiple physical servers which made scaling an application difficult.  Maintaining a coherent revision level control across applications of different servers was also a mammoth task.

Virtualisation has changed this.  Now the data centre resources, dedicated to an application, can be scaled up and down as quickly as an icon can be dragged and dropped on a screen and there is only one version of the application on which to maintain revision level control.  Now SaaS can offer a truly scalable service.  However, they still have one mountain left to climb – the trust element.  Will organisations really trust their critical applications and data to a third party?  

Today, many data centre’s are using virtualisation to make better use of their existing data centre resources.  But as these techniques become more trusted and the skill base to deploy them becomes more widespread, we may yet see the data centre itself becoming an internal SaaS – selling services back to the business units within the wider body corporate.

Virtualisation in the data centre supporting Service Oriented Architectures (SOAs)

One of the drivers for change in the application arena is the advent of Service Oriented Architectures – SOA.  This is where applications that have been traditionally stand-alone environments collaborate to provide a service to the business.  For instance, applications such as CRM, ERP etc. work together so that a single input from an application user creates interactions with other applications, in order to produce a single data set which is coherent across the organisation.  This approach requires support from within the data centre as applications must not only have a language with which to communicate, but also a network infrastructure which facilitates the interaction and the ability to seamlessly deploy the resources required to support these interactive applications.

While the challenge at the application level is to create a foundation for application interaction, this must be supported by a service oriented infrastructure, which using the network as a foundation, can deploy and redeploy virtualised data centre resources as required.  No longer can applications be seen as a standalone challenge, the whole data centre must now get involved if the opportunities and challenges of the future are to be addressed.

Consolidated Data Centres with remote application users

Consolidation into a single, headquarters-based data centre is becoming a trend.  This is being driven by the need to minimise duplication of IT resources in branch offices (both budgetary pressures and greener because resources are more efficiently used by sharing across remote office requirements) and to ensure that compliance and regulation requirements are met without having specialists at every location.

However, this raises a new challenge.  If all the significant IT assets (servers, applications, storage etc.) are consolidated into a centralised facility, how do we maintain the same speed of operation for staff at remote offices when applications are being run at the HQ data and access is over the relatively slow speed of the internet.

The answer comes from technologies such as market leader Riverbed, which allows remotely connected users to have the same performance from remote applications as they would have from locally connected servers.  This innovative approach employs a blend of techniques.  Some of the techniques used include reducing the number of handshakes required when a user accesses an application, caching techniques and reducing the volume of data movements by advanced data management techniques.  The net result is that the remote user enjoys the same speed of access to a remotely deployed application as a local user accessing that same application over the LAN.

This is a also a real win/win situation for the green data centre.  Consolidation gives better usage of IT assets and budget savings.  Remote working saves commutes from remote locations for workers, saving fuel, energy and time and the consequent pollution is eliminated.  As these techniques evolve, this can be expanded to include the home worker or telecommuter, making staff working from home a more realistic option for organisations and saving the pollution created by travel.  The real hurdle here will be to overcome corporate cultures which believe staff working at home will reduce productivity as they cannot be closely supervised.