Data centers are crucial to businesses’ IT operations, and in today’s society, are responsible for almost every aspect of BAU. Stakeholders, staff and end users all expect uptime all the time, and the best possible performance in anything they need to access when they want to access it.
Realistically, however, not many organizations have a team of highly skilled data center administrators available 24/7/365 to proactively manage and monitor all storage, server and networking operations to ensure that everything is working as it should – and this is before you consider security considerations, utilization constraints, maintenance, back-ups and replacements/repairs. But, there are some key areas to prioritize that are often overlooked; and their impact can have a knock-on effect on various factors.
At Procurri, we work with thousands of organizations of all shapes, sizes and types worldwide. This expertise has led us to formulate this list of the most common data center mistakes we see companies commit… and how best to avoid them.
Common Data Center Mistake #1: Poor Infrastructure Planning
Most businesses don’t have specialist staff internally who can design and build physical (or hybrid) data center infrastructure. As such, this job is often outsourced or handed over to a third party company – and unless properly and thoroughly briefed, they may only produce a data center to handle the organization’s current demands without future proofing.
Lack of capacity for Business Expansion
Often when a data center is being created, it is done so to a set brief and this does not include the potential growth or expansion of the business and its data and operational demands. The importance of sufficient storage space cannot be overstated, as insufficient capacity can result in system overload, downtime, and the need for emergency purchase and installation of further memory. This is expensive, time-consuming, and can deliver a poor service to end users.
When crafting data center infrastructure, it is key therefore to consider:
- Is there capacity for seasonal or ‘worst case’ demand?
- Is there capacity to meet the needs of any expected or projected business growth?
- Are planning regulations in place to allow expansion as/when required/desired?
- Is there sufficient physical space within the data center to allow for expansion as/when required/desired?
In order to avoid this, it is crucial that data center infrastructure is created not just with current requirements in mind, but also with future growth and development considered.
Thermal Management Challenges
Data centers are power-hungry by nature and so are prone to creating a lot of heat and hot air. It is now standard for data centers to have cooling and thermal management systems installed in place to mediate temperature and maintain ambient conditions.
If not properly maintained, data center hardware can overheat. This can cause disruption and downtime for end users, as well as irreparable damage to the hardware affected. There are a variety of thermal management solutions available for data centers, including water-based, airflow-based and refrigerant-based systems.
Critical cooling factors are considered and included in initial data center design. However, for those organizations running data centers in particularly warm geographic locations, it may be best to locate the infrastructure somewhere with the most possible cool air or cool water flow – often in the far north or far south of a territory, or nearby a natural source of water.
Changing (Physical) Space Constraints
While hardware, particularly in infrastructure settings where it is rarely thought of by anyone other than those managing it daily, is often thought of as installed and up and running, it doesn’t last forever and will require refresh after a few years.
When consolidating old and new servers, and in managing product refresh cycles, it should be noted that newer models of hardware are often considerably smaller in physical size than older ones, and so maintain a smaller physical footprint. However, this does not mean that they require less power – and often actually demand significantly more as they process more data and hold a higher digital footprint even where their physical has been cut. Such changes to physically smaller equipment is often thought of as requiring less cooling systems, but in reality need more and can even present challenges to the facility in terms of thermal footprint.
While it is, indeed, impossible to predict the exact future configuration of a data center, designers and administrators can maintain as much flexibility as possible to allow for changes as they occur. This permits the best possible chance of managing hardware recycles and adapting to them, rather than finding they create further challenges for the physical space and its functionality.
Common Data Center Mistake #2: Power Supply Issues
Data centers are notoriously energy-hungry and often require great deals of power. Despite the power supplies used in such environments usually being UPS (Uninterruptible Power Supply) facilities, they are uninterruptible in extenuating cases, and unplanned outages can cause vast downtime and disruption.
As batteries age, they can degrade in performance and if exposed to extreme temperatures (particularly overheating), can fail. Furthermore, if a power system is overloaded, it can shut down to avoid overheating and malfunction. This primarily happens when unexpected due to a lack of initial testing, and so can provide a sudden outage that causes disruption at the busiest possible time.
To avoid UPS or other power failures, proactive and regular maintenance and inspections of the data center should include power system monitoring and testing. Full loads can be tested quarterly or annually to verify capacity measures and constraints, and the use of N+1 or 2N setups for UPS, PDUs and generators help ensure that should one unit fail, power does not cease entirely.
Common Data Center Mistake #3: Networking and Cabling Constraints
Networking is one of the key portions of data center functionality, but there are two challenges often experienced by businesses in relation to their operations.
Bandwidth is critical to how much data can travel over a connection. Higher bandwidth means better speeds for the network as a whole, but this can result in higher latency too – that is, the speed that data takes to travel to its destination due to its increased distance. In order to manage this as effectively as possible, the physical cable structure between the server, storage and networking should be designed as cleverly as possible.
For cabling to work as effectively as possible, businesses should opt for high-quality certified cables only. Strict labelling and documentation should be filed on all and pathways between equipment should be structured to respect bend radii and to separate power and data sufficiently. Regular data center audits should include cabling inspections to help prevent onward issues. High-capacity cables such as Cat6a or Fiber can help future-proof by providing room for expansion.
Common Data Center Mistake #4: Insufficient Security
Data centers can present a myriad of security concerns, and it can often seem like there’s no end to what can go wrong. From the physical security of the center itself to access and data related issues, all bases must be covered to ensure maximum safety.
Studies show that about half of data breaches are caused by malicious attacks, such as:
- Phishing
- Compromised credentials
- Vulnerabilities in software (internal and third party)
- Cloud misconfiguration.
In today’s climate, legal and regulatory requirements vary from territory to territory, but there’s no reason why businesses shouldn’t go above and beyond with best practice to ensure maximum security. This should include a layered defence approach combining physical and cyber-security with vigorous access controls, staff training and strict incident response procedures. To avoid security breaches, data centers administrators should:
Physical Security
- Use multi-factor or biometric authentication for role-based access to the data center environment
- Implement internal CCTV systems and smart building systems for environmental monitoring
- Secure sites with physical deterrents such as CCTV, controlled entry points and fencing.
Cybersecurity
- Verify all access requests with Zero Trust Architecture
- Encrypt all data throughout its transit
- Isolate sensitive systems with network segmentation to prevent unnecessary or malicious lateral movement
- Use Intrusion Detection/Prevention Systems for monitoring of connections
- Maintain all devices with regular patching and updates.
Operational Security
- Allow minimum access needed for roles and responsibilities
- Ensure robust incident response and disaster recovery plans are developed and practised regularly
- Conduct regular security audits and penetration tests to identify weaknesses or threats
- Securely back-up all critical data regularly.
Staff Training
- Train staff in security measures such as suspicious emails, phishing links and physical threats
- Perform thorough background checks for any staff with high-level access or access to sensitive information.
Data Center Mistakes: Speak to the Experts
Procurri are the channel’s leading experts in data center management and maintenance – so we know a thing or two about how to avoid costly and common mistakes in such environments. No matter how complex your existing configuration, or if you’re just starting out and don’t know how to cover all bases moving forward, we can help you find the most secure and efficient method for your data processing and protection. Contact the team today for more information and benefit from our expertise with your equipment!
