Data center commissioning explained

Commissioning a data center is a crucial procedure that guarantees all your hardware, software, and infrastructure operations are in line with the Owner's Project Requirements (OPR) document - the functional requirements of a project and the expectations for its use.

Considering their size and complexity, data centers consume massive amounts of energy. In addition, they are the core of local area networks and many corporate systems operations. Many businesses depend on the uninterrupted operation of their systems, and stuttering workflows can be detrimental. For this reason, a data center’s failure can have catastrophic financial consequences. 

Therefore, its commissioning, design, and assembly are more demanding and complicated than commercial construction projects. In this post, we’ll focus on every key aspect of a data center: adequate access to power, a cooling system, a communications infrastructure, and all the other necessary elements that make for a successful commissioning program.

Optimal data center functioning: things to consider 

When commissioning data centers for proper operation, the following aspects are crucial:

Electrical supply

Typically, the most expensive factor in data center operations is energy power consumption. This is because uninterrupted power monitoring and control systems are crucial for efficiency and must be thoroughly reliable. Such kind of maintenance calls for a multi-sided approach in which we must take into account the following:

• The relationship between system design and redundancy, fault tolerance, failure control, and maintenance.
• Batteries, power cells, and similar devices used in uninterruptible power supplies (UPS), serving as bridging devices when switching from one power source to another.
• Backup power supplies or fuel generators that make for on-site backup power sources.
• Parallel power sources used to generate multiple power supplies that can take over in case of a failure.

Heat management

Data center processes require the continuous running of computer components. During operation, these warm up and give off a lot of heat. Regulating the temperature is vital to the proper functioning of the equipment and its longevity. 

This is why installing a reliable cooling system is another essential element of running an optimally-performing data center. When commissioning a cooling system for a data center, consider the following aspects:

• Operational scope of computer resources intended for deployment in the data center.
• The local climate, an often overlooked but important factor to bear in mind when configuring equipment. If atmospheric conditions and the environment provide "free cooling," it should be utilized accordingly.
• How density of the computing resources in server rack shelves affects the cooling architecture for optimal operation, be it direct rack-based heat flow systems or a standard row airflow (hot aisle/cold aisle) architecture.
• As server rack density increases, so does the demand for constant cooling. Without it,  temperatures in server racks may rise from 72 to 90 degrees Fahrenheit within 75 seconds due to the lack of heat dissipating surfaces.
• The relationship between system design and the prevention of overheating in terms of malfunction tolerance, failover switching, and redundancy of critical components.

Geographical placement

The location of a data center must be well-suited for its construction in every way. That said, the site-specific requirements for commissioning the construction of such a project are less critical than those for the effective operation of the mechanical and energy systems. 

Nevertheless, they also need to be kept in mind. These conditions are distinct in their characteristics and performance, which are influenced by the environment. When choosing a location for a data center, consider the following factors:

• Stability, accessibility, and affordability of local power sources, as many as the data center requires.
• Consistency and access to cost-effective network connections as the computing resources are useless to the systems that depend on them if the network fails.
• Availability and accessibility of mirrored computing resources in other coordinated data centers to take over the workload in the event of a failure in one data center, providing continuous support to a large number of important computer systems.
• Certain systems call for exceptionally low latency, which is why the length of the path that network traffic has to travel to reach end users must be optimal.
• Specific rules and obligations that certain regulatory agencies impose, such as the requirement that data is stored within a defined regional perimeter.
• Climate suitability for data center construction and cooling requirements.
• Stable geological and weather conditions.
• Frequency of flight routes.
• How stable the local political environment is.

Management systems

Management processes and automation determine the coordinated operation of facilities and efficient and timely communication within the system. These are the key components that must be coordinated between design and construction to achieve optimal performance:

• A universal interface for control systems that provides quick access to each system to retrieve comprehensive reports and detailed data, along with dashboards and critical operational statistics and alerts when necessary.
• The operational order that all systems follow.
• Techniques and strategies for creating management system software.
• The individual equipment vendors implement the protocol standards, then it is up to the installers, engineers, and technicians to program and configure them accordingly. Open and shared protocols do not necessarily ensure proper interaction.
• Hardware and software for energy monitoring systems.
• Software- and hardware-controlled cooling systems.
• Power supply for control systems: uninterruptible power sources must power the hardware for the network and control systems.
• Local control systems support accessibility and reliability. 
• The possibility of errors from both control system designers and when configuring the hardware as networking fundamentals competency is its own discipline. A networking engineer conducts careful planning of networking infrastructure. 
• The competence of the employees in dealing with the implemented control systems.

Physical security

Only authorized staff with the proper credentials should be allowed access to the data center. The system running on the data center servers determines whether government agencies and their laws and policies govern these access restrictions.

General

A few broad factors determine all operational requirements, including:

‍

• Funds: any increase in redundancy and reliability requires an increase in initial expenditure and possibly higher operating costs.
• Grade assessment: the Uptime Institute conveys the internationally considered benchmark for data center performance as follows: 

• The assessment's four levels establish standards for maintenance, power, cooling, and fault capabilities while fitting within a particular business function range.
• The data center becomes more reliable with each level, but the complexity and initial costs also increase.

The phases of data center commissioning

The difficulty with which facilities such as data centers are commissioned can differ in comparison with the commissioning of other buildings. Data center commissioning typically follows five phases (note that they can also be applied to other settings), as detailed further below.

These phases include all design, construction, and post-use operations. They also involve elements that take place primarily during the construction phase. 

Furthermore, a comprehensive commissioning program also needs to incorporate design undertakings as well. It is possible to combine the five-phase method with an additional, sixth phase, sometimes referred to as "phase zero," which includes design reviews, OPR development, and other early interactions.

The data center commissioning process is not set in stone and may vary with different commissioning agents. Possible variants of data center commissioning are:

• Examining all operational, failure, and maintenance situations.
• Testing every conceivable condition or circumstance through a series of tests and their iterations.
• Putting integrated systems to the test.
• Increased complexity of the tests.
• Fewer samples with higher witness proof ratings.

Commissioning phase one

The first phase involves reviewing the provided documentation and performing factory tests. Testing of equipment or system models under controlled conditions is referred to as "factory witness testing." This often takes place at the manufacturer's or supplier's facility. 

In such an environment, it is simpler and more cost-effective to detect and correct errors in the design and implementation of hardware and software. Additionally, it enables testing several different systems one by one while the data center is still in the making, which improves the effectiveness of the entire process.

Commissioning phase two

During the second phase of commissioning, on-site acceptance testing takes place. Here, the unit is examined to ascertain that it has a good storage plan prepared, is in pristine condition, and meets each design specification.

Commissioning phase three

This phase covers the installation review, also known as pre-functional testing or pre-inspection. This involves checking whether all pieces of equipment are properly installed.  

Contractors and commissioning engineers must examine the installation to ensure that it meets all design and operational requirements. During this inspection, the equipment is turned on for the first time to ensure correct and autonomous operation. Any equipment that fails the test will need to be repaired before being tested further.

Commissioning phase four

Functional testing is part of the fourth phase of the commissioning process. It is often performed for separate equipment and parts or items that are closely interconnected, such as a chiller, its cooling towers, and associated pumps. 

Every control loop is examined, while the performance and difference between actual and planned activities are evaluated. If necessary, changes can be made to setpoints. This stage will also reveal any operational problems. Notably, this phase includes more individuals, organizations, institutions, and disciplines working together than in other stages, at least as far as testing is concerned.

Commissioning phase five

The final part of the test and the last phase (aside from seasonal tests that have been missed previously) typically focuses on how the structure responds to major events such as an unexpected power outage. At this point, it is important to show that every system, including cooling, power, cooling, general HVAC, and so on, responds properly, promptly, and in tune with one another to prevent interruptions to data center operations. 

Finally, the overall performance of the system is verified concerning all planned requirements. The systems are put to function with different loads and in different operating modes to display the correct response to equipment failures and problems in service.

Simplify your data center commissioning project

Dataknox offers commissioning monitoring software for data centers so you can streamline your project, as well as save your time, money, and effort. Find out more here.

‍