Andrea Danti - Fotolia
As COVID-19 brings prolonged stay-at-home orders, organizations must find ways to continue data center operations and maintain infrastructure without on-premises access. To support the increase in remote work -- both on the IT team and across the organization -- reliable infrastructure is essential; main goals for organizations should include business continuity and downtime prevention.
"It is clear the crisis has only emphasized the need for uptime via redundant and reliable infrastructure and PDUs are considered a key part to the critical physical infrastructure," said Calvin Nicholson, senior director of product management at Server Technology.
Beyond redundant and reliable infrastructure, the rapid adoption of virtual and cloud-based technology highlights any infrastructure issues and maximizes available bandwidth. This requires admins to address network issues and connectivity problems for remote staff.
"Overall, during COVID-19, we're seeing businesses and individuals being pushed into a high adoption rate for all types of collaboration and communication technologies. As a result, there is a lot of scrutiny on the last mile to the home, beyond the main networks. We've certainly seen some disruptions with the remote workforce trying to work from home and having spotty connectivity," said Lee Kirby, co-founder and chairman of Salute Mission Critical.
Address sudden infrastructure scale up
Increased network, virtual machine and cloud technology use mean IT teams must ensure that the infrastructure they manage can handle the increased bandwidth. The ease of infrastructure deployment depends on which technologies an organization uses.
"Organizations that designed their infrastructure to be cloud-native and autoscaling are doing great. At the other end of the spectrum lies traditional on-premises software, like many SAP ERP customers, some of which run on Big Iron UNIX systems, which are difficult to scale without a purchase order and lead time," said John Appleby, CEO of Avantra.
Lack of resource predictability is another challenge for data center operations. Most businesses have a clear understanding of peaks and troughs for specific occasions, such as holiday sales period or over month-end. According to Appleby, a 25% resource contingency is enough to handle most businesses' resource peaks. Currently, there are peaks and troughs of 75% reduction or 750% increase in resources, which represents an astounding 30x variability between different markets and highlights the need for scalable infrastructure.
Organizations that use colocation can work with providers to expand bandwidth -- network or data center processing -- as needed. With this rapid resource scaleup, there's also required supplies and hardware to run the data center and protect admins.
"I think a lot of organizations did not appreciate the significance of their supply chain and where they would be getting their resources, including even simple things like sanitation equipment. If it was a regional issue, they could have gone out of region and got it shipped in. But because COVID-19 has hit on a global basis, this has caused a layer [of complexity] that people just didn't anticipate. That's why data center providers need to have a clear business continuity plan in place that they can implement during abnormal situations. Once they know who their suppliers are, they should look and see if they could diversify that quickly if needed," Kirby said.
Keep data center operations online
Many organizations now wonder how their admins can manage a data center they can't physically access. This causes IT teams to turn to remote infrastructure management.
"Many of our customers are not even letting people into the data centers, which really boosts the demand for our remote monitored, outlet switched and environmental monitoring solutions. This includes understanding if there are any temperature [requirements] and that the proper humidity levels are maintained," Nicholson said.
For data center providers that must have essential personnel on-site, top fixes include plans to cross-train admins and shift schedules. Cross-training allows admins to address issues they were previously unfamiliar with or unable to troubleshoot; this is essential for operations as providers must limit how many people they can have in the data center at once. To be successful, organizations must have effective training policies and onboarding procedures in place.
"During a time like this, there is certainly more risk for those who don't plan and are thus unable to execute at the speed that they need to. For example, if a data center provider needs to bring on new staff quickly, whether it's to perform critical IT maintenance or simply move equipment around, they need to have good onboarding procedures in place. Human error, through either bad procedures or bad training, all comes to light in this type of situation," Kirby said.
Define priorities for business continuity
To keep data centers online, IT teams must know how to prioritize specific operational tasks, such as preventive maintenance, hardware installation, facilities checks and upgrades.
According to an Evaluator Group report, planned data center infrastructure upgrades will largely be put on hold. However, 26% of participants stated social distancing requires new IT infrastructure and increased network costs, which presents IT teams with conflicting requirements.
Before any upgrades, it's essential that IT managers refer to or develop a business continuity plan. The larger issue is that most IT departments might not be equipped to handle inaccessible infrastructure.
Managers should develop a prioritization plan of what projects are necessary, what projects can be slightly delayed or what tasks can resume post-lockdown. Staffing also affects certain tasks, especially if they require more than one person to complete. Depending on project severity it may just have to wait, as increasing the amount of people in contact with each other increases risk of infection.
"Say a data center provider faces a choice between deploying technology to handle the loads that are coming in to service their clients or performing maintenance on a generator. During a crisis like COVID-19, they're probably going to deprioritize the generator maintenance in favor of bringing the load on, as long as that load doesn't jeopardize operations," Kirby said. "My advice to my fellow industry leaders is to look at your business requirements first and what is needed to keep the business running. Then look at what it takes to sustain the business as a second priority."