Data centre management part one: drivers and disruption

by | Nov 15, 2017 | Articles, Maintenance & Management

Data centre management and maintenance practices continue to evolve driven by factors including new business requirements and emerging software tools. There has been a lot of innovation in data centre design over the past 7 or so years. From the relentless focus on energy, capital and operational efficiencies of some hyperscale operators to flexible and agile micro-modular ‘data centres in a box’, innovation has been rife.

New forms of cooling, innovative rack designs, power distribution methodologies, mechanical and electrical (M&E) equipment and management software have all been refined and improved.

However, inventive design and construction is only half the story when it comes to data centres being fit for purpose and maintaining uptime. Just as the performance of a Formula One car depends on the driver and pit-crew as well as the car itself, the effectiveness of any data centre ultimately rests with how it is operated and maintained. Good management practice can often be overlooked if the data centre operator has not established the correct culture around operations and maintenance (O&M).Importance
There are a number of reasons why data centre owners and operators should develop a coherent and standardised approach to facilities management:

Uptime and availability
Operators invest huge amounts of capital to ensure that facilities are as resilient as possible. However, this is not a safeguard against poor management. Up to 70% of unplanned downtime is due to human error according to Uptime Institute. This is often the result of underdeveloped and unenforced O&M procedures and documents rather than isolated mistakes by individuals. (For more see our article on engineering forensics)

Efficiency
A significant proportion of the total cost of ownership of a facility across its lifecycle is operational rather than capital expenditure. A lot of costs are sunk in during construction and fit-out but there is also ample scope to realize efficiencies during the operational phase. For example, a one-degree change in cooling set points can result in a 2% gain in operational energy efficiency of a cooling system. The installation of high quality aisle containment may allow supply air temperatures to be increased from circa 18°C to circa 24°C and in some installations this will provide a circa 12% increase in energy efficiency.

Business drivers
The business model of commercial operators, including multitenant data centre (MTDCs) operators, depends on the ability to meet, and hopefully surpass, the service level agreement (SLAs) and key performance indicators (KPIs) of customers. To that extent, effective data centre management is critical to an MTDC operator’s business model. The same is true for facilities management (FM) services providers that will take-over the management of all aspects of a client’s data centre from security to basic IT hardware maintenance.

Staff development and retention
Effective data centre management also depends on having skilled and qualified staff. Uptime Institute, for example, recommends that one to two qualified staff is needed on-site at all times to support the safe operation of a Tier III or IV facility. Part of the operational regimen for any facility is tracking staff training requirements and professional qualifications. This is also linked to the requirement to bring in third-party contractors for specific management tasks. Training and development is also key to staff retention.

Innovation and change
The roadmap for effective data centre management is outlined in operations and maintenance (O&M) procedures and documents. However, looking more widely there is a broader set of drivers and specific trends that affect how data centres are designed and managed. These include:

Hybrid IT
The move from predominantly enterprise owned sites to hybrid IT models where organizations may maintain on-premises infrastructure, a presence in colocation as well as hosting workloads in private or public cloud has implications for data centre management. Under this hybrid model, data centre managers may be required to not only oversee on-premises equipment but also negotiate contracts and service levels with third-party providers. There may also be a requirement to develop monitoring and management software that can span all types of infrastructure to provide a holistic view of an organizations facilities and IT equipment.

New data centre types
The emergence of new data centre form factors such as prefabricated modular (PFM) designs including small-scale micro-modular data centres (MMDCs) will also require organisations to develop new data centre management practices. For example, some MMDCs, for example to support the Internet of Things (IoT) applications – may be deployed outside of conventional whitespace in generalist buildings. This could require data centre facilities procedures to be integrated with those for standard office space and other more generalist locations.

Data centre management software
The direction of travel in data centre management is towards greater automation and instrumentation. This could include the use of building management systems (BMS) or datacentre infrastructure management (DCIM tools to help with asset management, monitoring and control. There are also specific Computer Maintenance and Management (CMMS) tools that can automate some facilities management tasks.

Another category of tools, sometimes referred to a data centre operations management (DCOM), is also emerging and can help to automate the updating of O&M procedures and asset information. Developed by suppliers including MCIM and Icarus Ops these tools have a range of capabilities from maintenance scheduling and tracking to management of and access to documentation.

Looking longer term, a number of data centre technology suppliers have developed cloud-based data centre management software also known as data centre management as a service (DMaaS). These tools allow data centre equipment to be monitored in a similar way to on-premised DCIM. However data from equipment (owned by the supplier and the customer) can then be pooled and analysed in the cloud to enable features like predictive maintenance or be fed into the supplier’s R&D process for the development of new equipment. The end game could see more centralised monitoring and automated management of facilities with features such as sophisticated predictive maintenance reducing the need for manual on-site management. It’s also expected that artificial intelligence (AI) will be embedded into these tools to automate some data centre management tasks.

Some equipment makers are developing so-called ‘asset as a service’. This effectively means leasing equipment to customers bundled with remote monitoring and services such as preventative maintenance. The customer no longer owns the equipment or is responsible for its maintenance.

Lights-out management
The idea of a ‘lights-out’ data centre is not new, but it is evolving. It is based on the concept of using remote monitoring and management to reduce, or entirely replace, the need for dedicated on-site staff. Operators such as colocation provider EdgeConneX have integrated a lights-out approach into the fabric of its business. The approach has a number of benefits. For example, staff costs are kept to a minimum, facilities can also be effectively managed in edge locations where qualified staff may not be available. Couple this with the industry’s skill shortage and there are clear benefits. EdgeConneX has developed its own EdgeOS data centre management software to automate and standardise a variety of management processes.

Future direction
The trends and technologies covered in this article will continue to be developed and refined with significant long-term implications for data centre management and maintenance. It’s likely for example that the amount of direct human intervention in facilities will continue to decrease over time with greater use of remote monitoring and automatic control for both IT and facilities infrastructure. But there will still be a strong requirement for design-stage information to be integrated into operations procedures even if those operations are increasingly conducted remotely.

Part 2 of this series will look at specific developments around operations and maintenance (O&M) procedures and documents.

Future-tech provides a range of design and management services. If you would like to learn more about effective operations and maintenance practices please contact us at info@future-tech.co.uk