Maintaining two or more strategic vendors can make a lot of sense for many areas of IT infrastructure and operations (I&O).
This is particularly true in relatively commoditised markets such as PCs and low-end x86 servers, where lower prices — from maintaining a competitive environment between two suppliers — are likely to outweigh increased operational costs.
Except for the most proprietary platforms, servers are an area where a dual-vendor approach can be beneficial.
>See also: Gartner on the rebirth of outsourcing
This argument is strengthened in the midmarket, where x86-based servers are most likely to prevail, despite conventional wisdom suggesting smaller companies should work with fewer vendors to keep things simple to manage.
Outlined below are some of arguments for each approach.
Enhanced negotiation power. In a single-vendor relationship, the supplier has less incentive to lower its prices. Even when a single-vendor strategy is working well, it can be useful to invite additional vendors to quote for new technology needs.
Improved account support quality. Once a supplier perceives a client to be ‘locked in’, there is little incentive to provide superior levels of service. A vendor that feels vulnerable to another supplier, or wants to become more strategic to your business, will be more responsive to user needs.
Access to more technology choices. A single-vendor strategy limits users’ technology options. Even if this is not a current problem, it may become one in the future. A dual-vendor strategy enables a choice from a wider range of options, which in turn should ensure a user can deploy the best solution for its needs.
Better mitigation of risks. A dual-vendor approach helps to hedge the risks of a vendor going out of business, or scaling back product development, or of a supplier being acquired by another organisation.
Avoidance of lock-in. Becoming locked in to a proprietary environment — Unix, for example — can be particularly limiting because application code, software stacks and even storage may be specific to that environment. In comparison, users that standardise on Windows or Linux can benefit from near-universal software compatibility, which creates a more portable hardware environment.
Simpler administration, operations and training. Even for seemingly identical x86 servers able to run the identical software stack and OS build, the administration tools can differ greatly. The single-vendor approach offers the advantage of purchasing and learning to use a single set of tools and competencies. Hence, a single-vendor approach will lower operational costs at the risk of greater capital expense.
Administration staff preference. The preferences and prejudices of administration personnel help to maintain single-vendor loyalty, even when there are no workload factors to limit vendor choice.
Lower support and testing costs. When two vendors are used, repair personnel must be trained on two server types, and dual sets of spares may be required. When applications require certification on two sets of systems, testing costs are higher.
More reliable and efficient infrastructure. Single-vendor deployment lowers the probability of failures caused by incompatible software and hardware components. Testing and validation of new software and applications becomes faster and easier, and it is also easier to move people from one part of the IT organisation to another — because the local infrastructure does not vary.
Increased discount potential. Dealing with one vendor inevitably leads to a higher sales volume — so bargain for a better discount that might outweigh the price reductions in a competitive scenario. End-users can compensate for lost competitive pressure by exploiting the fiscal year of the sole vendor and bargaining lower prices for completing deals in a timeframe agreeable to the vendor.
Ability to leverage hardware and software synergy. Vendors in the hardware business that also have extensive software interests (e.g. Oracle and IBM) may be able to offer advantageous support, licensing or performance when their software is run on their hardware.
Closer vendor relationship. Having only one account team to deal with is more efficient and will be particularly valuable in large organisations with complex procurement procedures. A single, close relationship with the vendor account team can also bring human benefits. The supplier can become a partner with a vested interest in the success of each project — thus sharing the risk and the reward of close collaboration.
More than two vendors
There are also a few arguments favouring long-term deployment of three or more vendors.
Legacy infrastructure that is still satisfying business needs has to be maintained and this may often be at odds with organisational standards for new projects.
When a new software stack demands the use of a third hardware choice, there may be no choice but to comply. When the software is x86-based, always check that the support demands are genuine rather than channel partners being overly eager to win additional business.
Finally, hyperscale and high-performance computing workloads can be distributed across multiple node types because there is little dependency on value-added features from any one vendor.
This allows a choice between multiple top-tier vendors, and creates opportunities for smaller vendors and ‘white box’ vendors to penetrate the business.