As a reseller, managing the availability of your clients’ mission critical systems requires an understanding of the risks and costs of losing access to business critical information or services, balanced against the cost of achieving a certain level of availability.
That balance is shifting toward higher levels of availability as network services become essential to business continuity and the cost of downtime escalates. As a result, IT managers face growing pressure to drive network availability to unprecedented levels. At the same time, new technology categories, such as cloud and virtualisation, add to the complexity of this challenge.
This pressure is not limited to the data centre. For an increasing number of organisations, the network itself is mission critical. It may not always be possible for remote systems to achieve the same levels of availability as those in the data centre, but the gap can be closed by applying the strategies and technologies used in the data centre to systems outside it.
The Four Keys
Continuous availability of mission-critical systems rests not only on flawless operation of the systems themselves, but on the infrastructure that supports those systems. Achieving ‘five nines’ network availability requires the installation and management of an infrastructure which supports continuous availability. Typically this centres on four key components: mission critical power, mission critical cooling, monitoring and management, and proactive maintenance.
IT systems have two connections to the outside world: A connection to electrical power and a connection to the network. Without power, information technology becomes worthless. Unfortunately, the modern power grid does not adequately support IT operations – and it appears to be getting worse.
Electricity demand has jumped 30% over the past 10 years, while transmission capacity has increased only half that amount. Consequently, your role is to design a power infrastructure that can deliver the power quality and reliability your clients’ IT systems require. Accomplishing this involves selecting the right equipment (UPS) and system architecture (topology).
Cooling is also critical to ensuring network availability, and becomes even more critical as the trend toward server consolidation accelerates. Packing more processing power into a smaller space and packing more servers in each rack can test and even exceed the limits of raised-floor cooling.
Facilities have dealt with this situation by increasing rack spacing, essentially distributing the heat from the equipment over a larger space. This offers, at best, an interim solution. Increasing rack spacing quickly consumes available data centre space while reducing the number of racks the data centre can support.
Combining room-level precision cooling with supplemental high-density cooling offers the most economical and energy efficient solution.
3) Monitoring and management
Despite increasing pressure to boost network availability, support system management remains an under-used strategy for maintaining and elevating availability. By collecting data that can serve as the basis for a preventive maintenance programme and providing early warnings to the right people at the right time, monitoring solutions can help avoid disaster. At the least, they support faster recovery.
A basic approach to support system monitoring should cover mission critical power and cooling systems. As availability needs rise, monitoring can be extended to other support equipment, such as generators, automatic transfer switches, physical security systems and leak detection systems.
For systems operating outside the data centre, critical system management provides essential visibility and control. Communication cards in remote UPS systems provide a constant update of UPS status and enable remote restart of connected equipment. Temperature monitors can be installed in racks or small rooms to ensure heat does not reach dangerous levels. These inputs can be integrated into the centralised monitoring system in the data centre or into a network management system.
Maintaining support-system availability requires a well-coordinated preventive maintenance and emergency service strategy that addresses both mechanical and electrical systems.
Organisations which choose to handle maintenance themselves should work with support-system vendors to identify an effective preventive maintenance schedule and to obtain instructions for handling failures. They must also obtain a list of parts to keep on hand to aid prompt recovery in the event of a failure.
That inventory should be included in total ownership cost calculations. The wait time to receive parts not kept on hand should also go into evaluating the do-it-yourself approach.
A holistic approach
While many resellers tackle each of the above facets individually, a holistic approach is far more appropriate when it comes to selling network infrastructure. That’s because the degree of application required in each case very much depends on what you’re doing elsewhere.
By far the most important aspect of any modern data centre or computer room design today is energy efficiency, followed closely by availability. The holy grail is an infrastructure design optimised for maximum efficiency without compromising on availability, which is why a holistic approach is critical. Add to the mix the importance of real-time monitoring, ongoing management, system planning and design, and 24/7 service, and the picture becomes very clear.
The hard way is trying to do it all yourself. The easier way – find a supplier who thinks and works the same way, and has a clear understanding of how the seemingly disparate pieces of mission critical network infrastructure fit together.
If you keep your eye on the bigger picture, you need never lose sight of your core business.