Ongoing proper management needed to ensure efficiency and effectiveness
Data centre managers should re-evaluate some of their methodologies
DATA centres are often referred to as the backbone of an organisation, which is why when it comes to the management of these facilities, companies invest a large portion of their capital and expertise into their design.
However, once successfully installed, many organisations fall into the trap of neglecting to properly maintain the operations of the centre.
Not dissimilar to computer maintenance, data centres also need ongoing proper management to ensure their efficiency and effectiveness. Like most glitches and breakdowns, data centre neglect stems from two key factors: Lack of knowledge about a company’s business goals and future plans; and poor understanding of data centre operations.
For years, the data centre industry has accepted that human operational error is the No 1 cause of data centre downtime. The key in perfecting data centre management lies in the methodology applied behind the maintenance and operations programme.
In this article, the five common mistakes that data centre managers usually make will be addressed, along with reasons why they should re-evaluate some of their methodologies.
1) Using raised floors for cooling
As a lot of power is required to effectively cool a data centre – efficient power-distribution and maintenance of cool conditions is vital.
A common and even traditional method to cool a data centre is to install a raised floor during the design phase of your facility. This enables air to circulate around cabling and racks, limiting but not altogether avoiding the formation of hot spots, making it an ineffective holistic solution.
Raised floor cooling also limits the ability for your data centre to expand or cater for heavier loads, and also incur high maintenance as regular cleaning is needed to prevent debris and moisture from being trapped in the plenum.
More effective cooling requires a carefully thought out strategy and data centre design which needn’t be costly. Simple no-cost decisions made in the design phase of a data centre, as well as straightforward operational procedures, can be hugely successful at cutting expenditure.
Powering off unused equipment: In large data centres, powering off idle servers can make a dramatic difference in the temperature of the facility and in turn lessen the pressure of the air-conditioners cooling around that space.
Arrange racks to create hot aisles and cold aisles: This will cost nothing more than some time and effort but will most certainly reduce energy loss, and prolong the life of your servers.
Tune redundant systems: Redundant systems are those used to back up primary systems like a UPS (uninterruptable power supply), server or similar. As these are typically running at far below their rated capacity, data centre managers should tune them for fractional-load efficiency.
Install adaptable power and cooling: The notion is to ‘right-size’ your data centre so that you have enough cooling capacity to handle the IT load that is actually being run, not the load the data centre can handle when it’s at full capacity.
2) Hiring temporary or unqualified staff to man the data centre and assist in its operations
Human operational error is the main causes of data centre downtime, yet companies still commonly hire personnel who don’t have the necessary qualifications and training needed.
This is an expensive mistake given the very real money woes to compensate for high staff turnover, operational errors and potential equipment failure resulting from poor staff.
Data centre managers must consider the implications associated with quick-fix hires and understand the importance of skill and knowledge, particularly given the data centre’s technical environment.
Once a company has established a qualified team, it’s important to provide them with proper development and training. Positive working environments often boost employee retention rates and knowledge retention.
Training plans should be drafted to effectively provide training in a format that increases the level of expertise for all individuals according to their skill levels. Plans should also include mission-critical drills to enforce the right responses into the staff until it becomes almost ‘second nature’ to them.
3) Considering their data centre is strong enough to withstand risks or thinking potential disasters won’t happen
In Asia, news about natural disasters such as tsunamis, floods, and earthquakes are unfortunately common play. These disasters often lead to loss of power which results in:
Partial or complete facility shutdown;
Weak performance of data centre equipment; and
Potential loss of critical data
Instances like these can result in companies investing in costly data centre recovery and re-building. To help address this, it’s critical for data centre managers to implement emergency procedures for a prompt resolution of power problems while minimising the impact on critical data centre systems.
Trained maintenance staff should always be emergency-ready and available so an immediate response can be organised.
Another common risk management pitfall is an organisation’s failure to identify the long-term benefits of spending on a higher efficiency uninterruptible power supplies (UPSes).
While many are hesitant to fork out the initial substantial sum for a higher-efficiency model, the savings in energy and cost over time can quickly compensate the original cost premium. Assurance against losing critical data during emergencies and disasters is also increased with the presence of a higher efficiency UPS.
4) Saving costs by manually documenting data centre operations
Many organisations forgo software management tools in order to save costs which can be harmful.
Beyond the monetary commitment, virtualisation is a key step to the future, and a step that can be a more economical one in the long run.
Virtualisation saves space and reduces the redundancy of having multiple and dedicated hardware for individual tasks by centralising all functions into a single machine hosting virtual servers.
It is also one of the most significant drivers of Data Centre Infrastructure Management (DCIM), which is the base foundation required in stabilising the data centre facility.
In addition, DCIM provides end-users with greater control over their data environments and allows them to quickly identify hot spots, overloaded servers and the like, and adjust them accordingly, enhancing the facility’s efficiency.
5) Working with a ‘good for now’ mind-set
Failing to consider or recognise an organisation’s expansion and how this will impact the data centre is another common pitfall. This short-sightedness inevitably leads to extra costs when addressing expansion down the road, as well as the possible degeneration of data centre capability and efficiency once the supply fails to meet the demand.
A modular and scalable data centre approach – which refers to the build-out of data centre space and MEP (mechanical, electrical and plumbing) capacity in smaller increments – is an effective solution, allowing the data centre manager to scale upward in proportion to the end-user demand.
Preserve capital: Just in time delivery of needed space and capacity will delay and preserve capital deployment and increase the net present value of your total capital investment.
Reduce annual operating expenses: Minimising the amount of equipment installed proportionate to the current demand will reduce the amount of annual maintenance required. Additionally, it will allow the equipment to run at full load and operate at its peak efficiency, thereby reducing your annual energy consumption.
Prevent stranded capacity: The balance of power and cooling to the white space (watts/s.f.) is challenging. The trend has been to increase the design density in the white space. Often times this density never materialises and results in stranded power and cooling. Since the MEP infrastructure in a data centre can constitute 75% or more of the total construction cost, stranding power and cooling is undesirable. A scalable MEP plant will allow you to build to the density required within the available white space.
The data centre is the backbone of a company and it needs to be correctly managed. Design is one component of its success; the other is maintenance, which organisations cannot afford to address as an afterthought.
It is the data centre manager’s role to ensure they are aligned with their business’ goals and implement a strategy that considers both current and future plans.
Pankaj Sharma is the IT Business vice president for Schneider Electric East Asia, where he oversees the business strategy, ensuring that businesses across the region remain focused on customers and partners.
Data center issues hampering business strategy
Data centres in Malaysia and Singapore just too cool for words
I’ll get right on that, boss – data centre ownership
For more technology news and the latest updates, follow us on Twitter, LinkedIn or Like us on Facebook.