Andy Bolton, Chief Executive Officer, Capacitas, was interviewed by Drew Robb, freelance journalist in 2004, for Building a Model Infrastructure published in ComputerWorld. Material was also used for Capacity Planning: Art, Science or Magic published on EarthWeb.
Drew Robb: What are the latest technological developments in the capacity planning area? What can we expect to see in the near future?
Andy Bolton: The various platform vendors and service providers, rather than tools vendors, are currently leading capacity planning developments, predominantly with Capacity-on-Demand solutions. In telecommunications these products include Level 3 Communications’ ONTAP product, and in IT IBM’s e-business on demand service. There have been many new specialized tools that focus on a small part of the network stack, especially optimising SONET ring, IP, ATM or MPLS networks.
Drew Robb: How much do the tools help and how much is still
Andy Bolton: Any tool is only as good as the input data that drives it – they all follow the garbage-in-garbage-out philosophy closely. However there are many tools coming to market which deliver optimisation results that could not be achieved by hand. Generally most in-house tools are simply built to deliver automation benefits, rather than complex algorithmic solutions. Capacity Planning as an art form has become more important since the dot.com meltdown, as it has demonstrated the dangers of building networks to dubious over-rated forecasts or philosophical visions. Networks should only ever be built to customer demand, which is difficult but achievable. This demonstrates the importance of the capacity planner keeping their skill-set updated with new capacity planning and performance management innovations.
Drew Robb: How big of a role do they play in helping to accurately provision for forecasted needs?
Andy Bolton: The biggest downfall of most tools is getting the data in – getting accurate data is even more difficult. Data needs to be pulled from a variety of databases or OSS applications, often with propriety interfaces. In the worst case it must be manually entered, such as when defining a network, although many tools, such as WANDL’s products, can now automatically generate network maps from router data. Once accurate data is in the tool most can produce results reasonably quickly and within user-defined parameters.
Drew Robb: How big is the market? Is it growing? If so what drives that growth?
Andy Bolton: The market appears to have stagnated during the recent telecoms bust, although most vendors should be recognising a recovery by now.
Drew Robb: What should someone look for in selecting a capacity-planning product?
Andy Bolton: Customers looking for a capacity-planning product first need to be realistic about their requirements. No product can do everything required in a capacity planning function, so this fact needs to be accepted first. The customer should firstly make a comprehensive list of their requirements, both functional and non-functional, before even engaging in conversation with any tool vendor. These core requirements may miss some features that the customer didn’t realise were possible, but these other features are the optional requirements, not the core requirements.
Drew Robb: How well do they do their modelling? Are there products that model the entire layer 1-7 stack, or are they still split between L1-3 and 4-7 products?
Andy Bolton: Tools fall into several functional categories, which include: data collection, data representation, trending, analytical and simulation modelling. To my knowledge no tool currently plans all 7 layers of the network stack, as most concentrate on closely linked network layers, such as IP/ATM or IP/MPLS. Although this is a very fragmented approach it fits well with most carriers and corporations who build networks, as their network teams normally only cover a limited number of layers. Organisationally this makes getting value-for-money from a comprehensive multi-layer tool difficult without imposing a solution with all the pain of change that that induces. There appears to be a push towards convergence of network and application tools amongst some vendors.
Drew Robb: For most companies is there real value in using a capacity-planning tool, as opposed to just using the trending features on their management software to see when they will need to add capacity?
Andy Bolton: Trending works extremely well in a few limited circumstances – but unfortunately it doesn’t work as well as most of its proponents imagine. However it does enable reasonably accurate quick-and-dirty analyses to take place, with a good return on investment for this work. However the capacity buffer needed due to trending inaccuracies is often greater than the cost of performing more accurate capacity forecasting analysis – like all activities a cost/benefit analysis is needed for spending on capacity management tools, training and even staff. A mistake in capacity planning can lead to lost revenue, service remedies, stranded capital, lost customers or, in the worst case, failure of the business. These are high penalties to pay for not spending on a few analytical staff and tools.
Drew Robb: Anything else the readers should know about?
Andy Bolton: Capacity planning is one of the most important support functions in any company – the potential for disastrous mistakes is effectively infinite. Good staff, relevant training, and effective tools are equally important, as well as a recognition at senior management level of the need to invest in accurate analysis.
