When businesses move workloads off-premises, they must consider many important factors that impact their availability, scalability, reliability, security and other requirements. Companies that have dependencies on specific network configurations or non-standard operating systems often leverage colocation data centers. But within the colocation spectrum, you need to choose between human and automation managed data centers.
As data centers evolve and become more advanced, the resources and tools that managed them must do the same. This need is even more apparent with the rapid growth expected in traffic flowing through data centers. A recent study from Cisco predicted that global data center traffic would grow from 4.7 in 2015 to 15.3 ZB in 2020 – a 226% increase.
The approach to managing data centers and their substantial traffic is an important consideration. As the IT world moves from reactive monitoring to proactive management, disruptions will be fewer and farther between, issues should be resolved more expediently and efficiency should be much greater.
How are data centers attempting to be more proactive? How do human and automated data center management differ in their ability to prevent outages, maximize network performance and secure your data?
Research conducted by the Ponemon Institute discovered that the average cost of a data center outage was roughly $740,000 – a nearly 40% increase from 2010 and a per-minute cost of $7,900. What were some of the most damaging causes of data center outages?
The Ponemon study found that the failure of the Uninterruptible Power Supply (UPS) was the number one cause of data center outages – 25% of incidents. Tied for the second most common cause of unplanned downtime was DDoS attacks and human error, each accounting for 22% of all outages. A study conducted by Uptime Institute – the data center certification body – found that the greatest cause of human errors stemmed from oversights in staff training.
Errors of automation certainly occur as well. When too much is expected of a system or scalability is slow, outages can be the result. When that happens, it’s up to the Data Center Infrastructure Management (DCIM) system to repair and restart itself. This is a reality with which many IT leaders aren’t yet comfortable.
There’s assurance in having a real person to talk to when an outage occurs, a team of technicians who will personally repair an issue. In fact, Uptime Institute recommends that data centers enlist one or two qualified technicians to support data center operations at all times.
Availability is a non-negotiable delivery from data centers to their clients. Whether providers guarantee 99.999% reliability or 99.99999%, your SLA is a written agreement to that standard. But availability is just the tip of the iceberg for maintaining your network performance.
IT staff of human-led data centers are constantly reviewing alerts from monitoring tools. When issues remain unaddressed, the effects on your network performance can compound. Moving from one task to the next will always take time for the most skilled IT professionals. Data centers that rely on systems to multitask and solve network issues may do so more efficiently.
Network switches can automatically load balance and maintain bandwidth by routing network traffic. And DCIM systems can change performance parameters, network paths and security rules based on your predetermined schedules. If network traffic exceeds a predefined threshold, the DCIM system could temporarily increase available storage to avoid a system failure. Once the traffic falls below the threshold, the excess storage would be freed for other use.
Importantly, DCIM systems gather immense historical data, which may reveal patterns in your flow of network traffic. Such information helps data center managers analyze common trends and better optimize your network. Where software may struggle to solve large-scale problems that require solutions outside of typical operations, humans can learn from traffic data to creatively address recurring issues. Trend data also helps you design more informed capacity thresholds and plan for well-timed maintenance, which will ultimately minimize your data center costs.
While data security is critical to any business, your company may operate in an industry that requires strict compliance guidelines. You need to review any compliance concerns that may impact which data center you use.
As with managing network performance, manual systems that require human involvement to patch problems can be slow. From identification to fix, it could be a matter of hours, days or weeks before security vulnerabilities are addressed. As soon as security holes are discovered, automated systems immediately being working on implementing patches.
Machines can rely only on algorithms and data to process and address incidents, however. Only humans have the ability to make informed and immediate decisions. Critical thinking is a human characteristic and linchpin to devising the most-suitable solutions to security issues.
The prevalence of data center automation has grown over the years. With the speed and accuracy at which DCIM software can address issues, it’s reasonable that automated systems would become more common. Yet, thought leaders such as the Uptime Institute recommend human involvement for a reason.
While the most advanced data centers will rely on intelligent systems to manage part of their operations, expert and experienced data center professionals are key players in helping you achieve your business goals. From planning for optimized network performance and cost to supporting your business through any issues, human involvement is still a necessity.
Learn how to select a data center provider that can address your problems and add value to your network by downloading your copy of the informative whitepaper.