When it comes to the architecture of mission-critical systems for any organization, the power and network infrastructure layers need to be reliable, redundant and resilient. Those are the three R's you need to remember if you are working on building or maintaining any mission-critical application.
There should be no single point of failure in either one of these layers of intelligent infrastructure. If there is, you need to reconfigure the power and/or the network infrastructure to eliminate that disaster waiting to happen. Single points of failure should be totally eliminated in any facility supporting mission-critical applications.
Think about the building or business campus your customer’s company or organization is housed in. Chances are, it has a single connection to one central office for its broadband connectivity requirements and a single connection onto a single power grid for its power requirements.
The single connection to a single utility dates back to the horse-and-buggy days when the power and network management layers of the infrastructure were just being installed. Times have changed as well as the criticality of the dependence on those intelligent infrastructure amenities.
Facilities cannot support mission-critical applications unless they have power coming from diversified sources (two different grids) and communication systems getting their feeds from more than one network carrier coming down on diversified paths and not the same one from one central office.
Customers need to understand the three R's
How do you sell that to the customer when they tell you they are on a tight budget? They need to understand the importance of having redundancy and reliability when it comes to supporting any mission-critical applications so they can be resilient. When they say, “we are on a tight budget” you need to counter with, “How much of an outage can your business withstand if you lose any of these critical applications for an extended period of time?”
That is really a good management exercise for your customers to go through. If you are working on a data center or a call center, what is an acceptable outage? Thirty minutes? Three hours? Three minutes?
I worked with some retail corporate call centers where they could tell you down to the penny what each call was worth to them. They would know exactly what an hour downtime was worth. They could break sales revenues down from so many dollars per hour down to what is an average phone call worth to them in sales in both regular and peak (holiday) season business cycles.
If your customer cannot come up with those solid numbers, they are not managing that call center properly. You cannot manage what you do not measure. If this is the case, they first need to analyze the amount of business coming through the call center and figure out the amount of sales in dollars per month, week, day, hour, call, and down to each individual agent.
Once that happens, they can then compare what the downtime costs are versus what it would cost to build out any areas of weaknesses (improving the redundancy and reliability) of the facility.
The growth of mission-critical applications
Did you know one out of every three applications are considered mission critical across all organizations? That is a significant number and the supporting intelligent infrastructure for those applications need to be robust in both their design and operation.
This is a growing area. It is predicted it will expand to one out-of-every-two applications becoming mission critical, meaning intelligent infrastructure supporting it must be fully redundant and not have any single point of failure.
Electrical contractors should look at this as a new opportunity to build out the layers of intelligent infrastructure focusing on redundancy. Part of any facility review should be the quality of its supporting cabling infrastructure to the business and its applications.
When it comes to working in data centers and building out their network infrastructure as well as their power requirements, remember, best practices are a moving target.
Best practices can also change with the weather. What was state-of-the-art a year ago, could be totally obsolete today.
Understanding the need to monitor systems
Cyberattacks, cyberterrorism and natural disasters cannot be anticipated as far as if, and when, they will occur. They can and will occur anytime and usually when someone is not anticipating anything except another routine day.
When they do occur, are you even aware that something just happened? Chances are, you may not be. I will discuss this in part two next month.
Editor’s Note: Carlini will address the IWCS International Cable & Connectivity Symposium in Providence, RI (in October) on this critical and timely topic.
Carlini is currently writing, Nanokrieg: Beyond Blitzkrieg, a book on military and critical infrastructure, strategies and tactics to fight the war on terrorism in the 21st century. It is planned to be out at the end of the year.