Network Attached Processing Feature Story

Handle Unexpected Requirements and Scale Your Business

Let’s face it. The last thing any enterprise IT organization is going to do is have ‘just enough’ compute capacity to meet ‘minimum’ business requirements. Not in this lifetime. You need to scale with the business, and you need enough spare capacity to handle unexpected requirements. So what are the choices? The traditional horizontal/scale-out or vertical/scale-up scaling models have been around for a long time and there’s good and bad associated with each of these axes.

Commodity Conundrum

The commodity building-block approach of horizontal scale is logical, but it is hard to determine how much power you need when and where? And because it is “distributed capacity,” you cannot share and minimize the spare capacity in each server. The result, 100% of the time, is to over-provision well beyond what you actually need to cover peaks. And as the number of building blocks increases, so does the complexity of managing the overall complex. Too many moving parts cost more. This has led to a tremendous amount of cost, complexity and power consumption challenges in today’s enterprise data centers.

Vertical Vertigo

The alternative to horizontal scale is to scale vertically. Use big iron, classic large-scale SMP systems. However, these systems are very expensive to buy and operate. Trying to isolate applications in an SMP environment can be difficult, especially when dealing with ‘moves, adds, and changes.’

Innovation Gap

The result is what savvy CIOs refer to as the “innovation gap.” Thanks to escalating capital expenditures, power consumption, cooling costs, real estate constraints, software licensing and more, the cost of keeping status quo is increasing rapidly while the total IT budget is staying constant or decreasing and the demands on IT are growing. It’s because of these unsustainable operating costs that less and less variable spend can be put towards innovative initiatives that can drive competitive advantage.

Something Has to Change

Traditional models offer no way out. And businesses are not about to reduce their need for higher ROI from their IT investments, demand new applications and services, and grow their user base. Neither the scale-out nor the scale-up model is able to deliver the value that is now expected. And it is no longer sufficient to simply reduce the purchase price, IT deployments must also be economical to operate, evolve, and retire.

The Wrong Hardware Increases Costs Everywhere

While hardware is a small fraction of the overall IT cost, it has a large and long-term impact on all other costs. Examples of costs that hardware systems can cause or avoid include: asset utilization, change management, meeting heightened end user expectations for instant response time, consistent quality of service, rapid deployment of new applications and services, and frequency and size of future upgrades.