CIOs have a tough job delivering an expanding range of services to ever more demanding customers and responding rapidly to requests. The challenge is made much more demanding as most organisations are running a disparate array of systems from multiple vendors, with each typically supporting a single application.
When looking at the issues facing CIOs, it is important to distinguish exactly how the challenges are perceived and dealt with. By this we mean explicitly separating out and recognising the root cause of underlying problems, rather than only noticing the symptoms these problems cause. With this in mind, surveys carried out by Freeform Dynamics highlight a number of issues experienced keeping IT infrastructures operating to meet business needs.
The results consistently show that many of the visible pain points, such as juggling priorities, IT staff overstretch and keeping up with demands from users often reflect deeper fundamental issues in systems, especially around the lack of integration between different systems and especially in the tools used to administer them. And on the people side, four out of five larger organisations have completely separate teams looking after server, storage and networking.
Vendors, sensing sales opportunities, have spent considerable efforts in recent years encouraging organisations to consolidate their servers and, more recently, their storage platforms, often making use of virtualisation tools.
Such programmes have delivered considerable benefits, especially in terms of resource utilisation and availability, with consequential cost savings, but in many instances have resulted in new virtualised server and storage siloes appearing, albeit ones larger in scale and now running multiple services. But reports from organisations that have virtualised significant portions of their estates indicate that operational management remains a challenge, most notably because of the many tools that must be employed to keep systems running and the fact that many IT departments split responsibility between separate teams.
In fact the ability to manage large scale virtualised platforms is a significant challenge when using multiple tools to administer each facet of the stack in isolation. Even worse, we know that many organisations find it very difficult to make a business case for investing in integrated IT management tools as few budget approvers are explicitly aware of the close linkage between business operations and effective IT managements.
To address these challenges, a number of vendors, such as Cisco, IBM, DELL, HP, Fujitsu and Oracle, have been pushing the concept of Unified Infrastructure. This involves combining server, storage and networking resources, all pre-integrated into a single platform and surrounded by a management interface to make daily operational administration more straightforward. In many respects these solutions have more than a passing resemblance to the original mainframe systems in that they can be administered as a single entity by teams responsible for the delivery of the entire service, not dependent on a server admin, a storage guru and a networking maestro.
To achieve this vision of Unified Infrastructure, many vendors have developed the ‘datacentre in a box’ concept, but are these platforms the answer to IT infrastructure challenges?
Such systems do offer many benefits, most notably the pre-integration of server, storage and networks thereby significantly limiting the amount of installation work required. Indeed, several vendors claim that their systems can be moved into the datacentre / computer room, be plugged into power and networks and be up and ready to serve users within a few hours rather than days. The tight coupling of components can also reduce the amount of testing required as operating system and firmware updates are brought out as much of this work should be undertaken by the supplier.
It should be recognised that most such systems are still some way from being off the shelf items. Indeed, most such solutions today are configured and assembled when orders are placed. And typically, these systems tend to be configured and specified as a result of significant consultation making the entire order – delivery process consume weeks, if not months for the largest of solutions. It is only very recently that one or two vendors are seeking to provide application or service templates designed to sit on top of preconfigured ‘stock’ items to deliver a total package more quickly, but these still have some way to go to reach maturity for a wide range of applications.
One clear advantage worth recognising is that such solutions are supported by a single vendor, potentially reducing the time required to service systems and fix any problems that might arise, a decidedly non-trivial matter in more complex systems.
There are also other factors to consider, most notably that the current range of offerings are not currently suited to small scale operations as the platforms tend to be relatively large and therefore beyond the needs, and budgets, of most small businesses. Until vendors produce small systems, pre-loaded with business applications, few small businesses will be able to take advantage. Even larger organisations often seek smaller footprint solutions to meet either specific requirements or as a result of budget limitations.
But perhaps the most pressing technical challenges concern the age-old bug bear of application and data integration alongside the problem of managing systems that run twenty-four by seven over extended periods of time. Clearly these challenges are not eliminated in the data centre in a box solution.
When it comes to data integration, we know that organisations are making more use of ‘composite’ applications which have functionality running on multiple systems and with data spread across many repositories. Getting the right data to the right application / user at the right time still needs some manual labour on set up and with each software upgrade and patch.
On the management side, organisations may need to ensure they have in place tools that will allow them to orchestrate and administer systems from several suppliers, if only to address fears of vendor lock in. More importantly, only with such capabilities will IT be ready to take advantage of new offerings as they develop and mature, a factor that may feature more strongly as external and internal cloud solutions bed in.
Beyond this, it is likely that many CIOs may have to alter the way their IT budgets function. Many organisations invest in compute, storage and networking in direct response to particular application service delivery requirements. Few CIOs find it straight forward to invest in shared IT infrastructure. Such an alteration may also necessitate bringing in some form of resource usage charging, or at least reconciliation to ensure all user departments feel they are not being penalised. But getting such alterations to the way IT budgets are allocated and reported takes time and considerable political will power.
Datacentres in a box will be useful in many scenarios, but they are not yet the answer to all problems for everyone. There are clear benefits to be had in running IT services on a shared infrastructure, especially if the organisation is ready to adopt a ‘service delivery’ model for running IT rather than a traditional ‘Infrastructure Operations’ approach. But it is important that any initiatives to build such systems do not result in larger proprietary silos rather than an open, flexible environment that can be managed by mutli-skilled teams.