Best of breed versus integrated multi-function solutions – it’s a discussion that’s been going on in IT for about 3 decades. Whether the debates occur in the context of software suites, hardware/software appliances, or complete pre-engineered system stacks, there are always vigorous advocates arguing the case for each approach.
One such debate surfaced in a recent Freeform Dynamics research project around ‘application delivery controllers’ (ADCs). These bring together elements of network security, load balancing, performance control and other related capability to produce multi-function boxes that can sit in your network and, well, handle multiple functions.
A number of participants in the research expressed strong opinions against the concept of ADCs. Familiar arguments surfaced around multi-function solutions always representing a compromise, only being as strong as the weakest component, and representing an unnecessary single point of failure. All good points, though it was clear that most of the naysayers didn’t have any direct experience of ADCs.
Fortunately, about a quarter of respondents were actually using ADCs at the moment to a significant degree in their network. This gave us the opportunity to compare and contrast the opinions of those with and without direct exposure to the technology (Figure 1).
As you can see, all of the familiar arguments for multi-function solutions come through quite strongly from those with experience, including ease of implementation, ease of management, simplification, and a lower ongoing cost of ownership.
In terms of actually doing the job, it is also telling that the most prominent benefit highlighted by ADC users is a lower risk of things falling through the cracks. This makes sense when we consider that elsewhere in the research, many respondents told us that their network included a lot of older components that had significant functionality gaps and were also hard to integrate and manage.
Of course none of these arguments will cut any ice with those who have a philosophical or religious objection to multi-function solutions, or a genuine need to mix and match specific components because their requirements demand it. But if a rapid ADC implementation is the difference between modernising and simplifying, or living with the gaps, disjoints and operational overhead in your network, then it could be an option worth considering.
And if you look at exploring the ADC route, it is worth noting that some vendors are delivering this kind of technology as a virtual appliance rather than a chunk of hardware. The majority of those that have an opinion on this say it’s a good idea that can represent a better fit with their broader virtualised infrastructure (Figure 2).
What’s also evident from this chart is the potential role of virtual ADC appliances in a cloud computing context, where you are spinning up hosted servers to run demanding and/or internet-facing applications. In the worlds of traditional hosting and Infrastructure as a Service (IaaS), providers are usually just giving you raw server resources. A simple drop-in solution to handle load-balancing and security can therefore save a lot of time and hassle.
Zooming out, ADCs are an example of a technology that has matured considerably over fairly short period of time. If we had conducted this research even 3 years ago, the picture would have been quite different, with frequent reports of extortionately priced solutions that were a nightmare to set up and manage and only suitable for high-end requirements. If that reflects your own view of ADCs, we would encourage you to look again, especially if you, like many others, have work to do in simplifying and modernising network infrastructure.
Pre-integrated multi-function solutions are definitely not right for everyone, but do seem to offer advantages in many situations if this latest study is anything to go by. If you are interested in seeing more of the research, which covered a lot more than ADCs, the report is available free of charge and without registration from here.