Driverless cars, robotic workers, implantable mobile phones, cryptocurrency payments, space tourism that’s the kind of world we envisage in 2030, a dystopian future similar to what has seen in sci-fi films.
Regardless of what happens in the years to come, one thing is certain: data centers will continue to play a vital role in storing, processing, transmitting, and managing information. The more data we generate, the more important our data centers will become.
In order to catch up with the ever-changing technological world, the data centers will need to evolve at a rapid pace.
So, what changes do we expect to unfold in the data center industry?
How will data centers keep pace with technological evolution while adhering to environmental regulations?
Will technological breakthroughs give rise to new forms of data centers?
Let us dig deep and find out.
On this page
Edge Computing
With the Internet of Things (IoT) getting more mainstream, IoT devices and sensors have proliferated, putting a strain on network bandwidth. Data centers are trying to take some load off the centralized servers by creating small, distributed data centers called as edge data centers that offer hyper-local storage and processing capacity.
In a traditional cloud computing model, the data generated by devices travels back to the centralized repository where it gets processed. Because the central server is situated far from the source of data generation, the data doesn’t get processed instantaneously. This delay brings some latency in the network that slows down devices.
Edge data centers are located close to the source of data generation (e.g. sensors, motors, generators), the so-called ‘edge’. These centers process critical data locally in near real-time; as a result, the flow of data to the centralized computing location gets reduced considerably. Edge computing, thus, results in more effective utilization of bandwidth.
With the IoT technology gaining ground, edge computing will become more viable than ever. According to a report by CB Insights, the edge computing market is expected to reach $34B by 2023.
Also Read: Beyond the Cloud: 5 Startups Bringing Data Centers to the Edge
Hyperscale Data Centers
As enterprises are increasingly deploying cloud environments, the need to offer high processing power has escalated. A hyperscale data center is a mega-sized center with hundreds and thousands of servers interconnected through a high-speed network. A hyperscale data center has a robust architecture with flexible memory, networking, and storage capabilities.
Industry titans Facebook, Apple and Microsoft all are building mega data centers to buttress their growing data storage and processing needs.
Facebook is building a gigantic data center spread across 2.5 million square feet in Fort Worth, Texas. Likewise, Microsoft is creating a data center cluster in Iowa. The cluster is spread across 3.2 million square feet; its largest data center, Project Osmium, built on 200 acres of space, is expected to be finished by 2022.
According to a report by Synergy Research, close to 430 hyperscale facilities are operational all across the globe. Another 132 hyperscale centers are currently in the development stage; when finished they will serve the needs of technology giants-Google, Facebook, Amazon, and Microsoft.
Innovative Cooling Methods
Data centers have traditionally relied on air-conditioning to cool their infrastructure. But air cooling has been considered an inefficient and expensive technique-it accounts for around 40% of the energy consumption of the data centers.
In a bid to create more energy-efficient facilities, data centers are experimenting with innovative cooling techniques. Artificial intelligence seems to offer a viable solution here: AI-based solutions can dynamically monitor and regulate the environment of data centers.
Google has recently tested an AI program created by DeepMind to regulate cooling in one of its hyperscale data centers and the results have been promising.
Liquid cooling technology has been touted as another viable alternative. In water-based liquid cooling, cold water is funneled through pipes situated along with the servers so as to bring down their temperature.
A variant of liquid cooling is immersion cooling, where the servers are immersed directly in a liquid coolant. In this case, the coolant is not water, but a dielectric liquid that does not conduct electricity.
Liquid cooling technology is expected to gain traction in the years to come.
According to a recent survey by HTF Market Intelligence, the global market for liquid cooling will reach $4.55 billion by 2023.
Hybrid Cloud Architecture
When cloud platforms became popular initially, several companies migrated their entire network to the cloud. While some companies benefitted greatly from cloud solutions, others had security concerns with the public cloud. As a result, some companies left the cloud altogether and switched to alternatives such as colocation.
However, many companies still wanted the convenience of a public cloud along with the security of a private cloud. Keeping in mind their needs, data centers developed hybrid cloud architectures that allowed clients to enjoy the combined benefits of the private and public cloud.
Hybrid cloud architecture allows enterprises to store sensitive, critical data in a private cloud while running applications in a public cloud. With a hybrid cloud, companies can now deploy public cloud for most of the applications and use expensive cloud resources only when needed.
SSDs and All-Flash Arrays
SSDs and all-flash arrays are increasingly replacing the structured SAN and NAS systems in data centers. All-flash arrays are storage solutions that have flash-memory drives instead of the traditionally used spinning disc drives.
Flash drives offer numerous advantages over mechanical drives: no moving parts, better IOPS, faster response time, less power consumption, and lower heat generation.
A major reason flash is replacing the conventional hard disk drives is performance. Flash drives offer better application performance in terms of IOPS and millisecond latency. The performance comes at a cost, though: flash drives are around 10 times costlier than their HDD counterparts. Plus, they have a lower storage capacity than hard disc drives; the gap on both counts is narrowing gradually.
According to an estimate by the IDC, flash drives are used for around 9% of the enterprise storage needs; the number is estimated to reach 15% by 2021.
Read: Exploring the Impact of AI in the Data Center
Enhanced Security
With data center in cloud computing, IoT and service-oriented architecture gaining ground, data center operations have become more dynamic, escalating the threat of security breaches.
Data centers are investing heavily in physical security infrastructure such as biometric system access, surveillance systems and monitored alarms. However, certain areas continue to be neglected.
An area of concern for many data centers is the inadequate monitoring of third-party personnel accessing the server area and the increasingly digitized operations they carry out within the premises. Datacenter devices are often connected to external networks that can jeopardize their security.
To counter such threats, many data centers have adopted a zero-trust approach. The zero-trust approach treats every device, transaction, and user as suspicious. So, no one from inside or outside the data center is trusted by default; anyone trying to access the data center network needs to verify their identity.
In the days to come, data centers are likely to adopt a conditional-access policy where they would allow access to a resource only if it fulfills certain criteria, including the time of the day, access location, the device used, etc. If these criteria are not satisfied or if some deviant behavior is observed, the access would be revoked.
Artificial Intelligence (AI) for Smarter Data Centers
Artificial intelligence, if deployed strategically, can revolutionize data center facilities and create smarter, more efficient data centers.
Gartner has predicted that more than 30% of the data centers that fail to prepare for AI and machine learning will not remain economically viable by 2020.
Organizations have begun to leverage artificial intelligence to streamline data center operations in several ways: optimization of workload distribution, creation of energy-efficient cooling solutions (discussed above) and mitigation of workforce shortage.
Traditionally, IT professionals have been responsible for optimizing the performance of their companies’ servers. The IT teams need to ensure the workloads get strategically distributed across various data centers. Constrained by resources, these teams often find it difficult to monitor workload distribution all round the clock.
AI-powered predictive analytics tools are being deployed to monitor and analyze the workflow of companies in an attempt to optimize their storage and computer load balancing in real-time. As a result of these tools, IT personnel can anticipate server demand even before a request is raised. Since AI algorithms improve on their own, the servers automatically become more efficient over time.
AI is being used to automating a range of IT management functions. AI tools can perform mundane tasks such as updating systems, security patching and backing up files, leaving more complex tasks for the IT personnel. As a result, IT professionals just need to oversee the routine tasks that previously kept them occupied all the time, allowing them more time to focus on the bigger picture.
Wrapping it up
The factors described above have been instrumental in revolutionizing the data center industry, paving the way for smarter, efficient data center services. The coming decade will witness more such developments driving innovation to an altogether different level.
Image source: Freepik Premium
