Data centers continue to drive the digital economy, either within enterprise IT or as part of various colocated or cloud solutions. But as data centers evolve, there are more types, designs and components to consider in any buying decision as businesses look to manage growing volumes of data and application usage.
In order to help you get the most value out of your data center and make the right decisions, this detailed guide covers the following topics:
- The evolution of data centers
- Understanding the modern data center
- The importance of data centers
- The core components of data center infrastructure
- How do data centers operate?
- Data center tiers explained
- The different types of data centers
- What is the future of the data center?
- How DCIM can help data center administrators
- Final thoughts
The evolution of data centers
First, there were standalone computers, then networks and servers, all helping businesses operate more efficiently as digital information became the currency of success. Over the decades, as the volume of information grew and the need to process it faster for a booming number of business users rose, data centers became the de facto cost-efficient way for enterprises to handle the load.
But as speed and storage increased, so did complexity and cost. Many enterprises get involved in costly data center migration operations, while most firms can’t afford a data center but require the power they offer. As a result they adopt colocated data centers provided by third parties, or do all their business in the cloud, relying on the tech giants’ own data centers to handle the load.
The rise in cloud and virtualization reduces costs, resolves issues like disaster recovery and redundancy and helps with security – all costly problems for any business to solve on their own. With 2,670 data centers in the US (the most in the world) and 450 data centers in the UK, according to Data Centre Magazine, these physical facilities power the telecoms giants, banks, enterprises and millions of smaller firms.
Understanding the modern data center
Given the costs and challenges of adopting and using data centers, the infrastructure has moved from the typical on-premises installations to virtualization and cloud solutions.
Wherever they’re located, data centers still maintain the same hardware, with racks of servers, modern CPUs and GPUs for processing big data loads and AI tasks, high-speed networks and security applications with fast storage required. These are supported by advanced power and cooling infrastructure, often with green energy commitments and managed and operated from modern dashboards, reducing the overheads and maximizing efficiency.
Today’s modern data centers provide businesses massive pools of resources and infrastructure, often shared among companies, providing virtualized data centers. They support their applications and workloads, with the business having to worry about upgrading to the latest processors, getting faster storage and meeting the enormous energy requirements of keeping these systems operational 24/7.
Going multi-cloud enables businesses to get the power it needs without the exorbitant cost of a traditional data center. They also offload many critical issues that data center operators encounter, including the ever-pressing need to provide greener operations, maintaining uptime, especially as climate uncertainty events hit energy grids, and the cadence of upgrades to ever more efficient components continues.
When it comes to choosing a data center provider, there are many factors to consider and questions to ask regarding redundancy, data accessibility, monitoring and other issues.
The importance of data centers
Gone are the days when a business had a few people in one office with computers. Now, every office worker, function and department is networked, while factory and field operators work on smartphones, tablets or notebooks. Sales and retail staff work in connected digital environments, and even delivery drivers move to a digital schedule. Also, business partners, supply chain and other valuable operations all provide digital data that must be processed, analyzed and monitored for anomalies and insights.
That volume of data and the services to process it rapidly moves beyond most local capabilities, especially organizations that handle large numbers of customers or other transactions. As storage and processing demands accelerate, the importance of data centers becomes clear.
Businesses need their capabilities but many don’t want the expense and time investment in building and operating them. This is why colocation or cloud data centers are increasingly the choice of high-growth businesses that need to scale fast and invest in business needs, not IT.
The core components of data center infrastructure
Data centers are effectively glorified servers for a business with added infrastructure and power, but there’s a broad series of terms and technologies used in marketing to make them appear special and distinctive.
Primarily, data centers consist of computing infrastructure, processors (like the latest Intel Xeons), storage (data center optimized SSDs), memory and network connections that link every worker to their email, spreadsheets, big data pools and local or cloud applications.
Providing connectivity is fast and there’s redundant network infrastructure, data centers are capable of supporting large numbers of users simultaneously with strong security. This is backed up by hardware or software solutions that ensure cybersecurity is tight through firewalls and malware monitoring, access is monitored and limited to valid users and compliance rules are followed.
What keeps everything running are the power and cooling systems, with various designs (hot/cold, high performance vs energy efficient) and ways of running data centers. These buildings are often highly secure, with limited access, multiple redundant power sources and other features to keep them running through typical outages that any city faces.
Take a look inside one of the biggest data centers to see what all that hardware looks like.
Note the very neat wiring, the energy and cooling solutions and the staggering amounts of information they can hold. All that technology can be monitored from a single room, tracking tens of thousands of data points simultaneously to look for anomalies or failures.
How do data centers operate?
Whatever the type of data center – broken down into tiers below – they operate in a similar manner. Internal data centers directly service the IT needs of their own business, while colocated or other types provide neatly-packaged data center services for a range of customers.
Contracts for colocation are more involved than a simple cloud service contract, with higher service-level agreements. The cancellation of a contract is more complex than dropping a cloud service.
Data center tiers explained
Tier 1 (Basic capacity): These are the most basic type of data center, with only one primary power supply, limited options for redundancy and backups. This reduces costs and management time, but are more likely to suffer downtime and extended outages if there’s a major issue, with estimates at around 99.6% uptime thanks to the robust nature of modern data center components.
Tier 2 (Redundant capacity): Adding more complexity in the form of redundancy and backup capabilities, Tier 2 data centers up the average uptime to 99.75% and are easier to recover from downtime issues, but they still typically have a single source of power.
Tier 3 (Concurrently maintainable): Improving on redundancy further, Tier 3 data centers offer multiple sources of power to avoid the most likely cause of an outage, with additional cooling and support systems in place to update and maintain operations. These increase the expected uptime to around 99.98%, so they could be down for only a couple of hours per year.
Tier 4 (Fault tolerant): The highest level of data center, Tier 4 offers high levels of redundancy features in all areas for totally fault-tolerant systems that can continue operations in all but the most unusual of circumstances. They offer 99.99% uptime, but naturally cost more to build, manage and license.
The different types of data centers
Data centers continue to be rolled out in more operations worldwide, with some 8,000 globally as of early 2021.
According to a US International Trade Commission report, a significant rise in data generation (data creation was approximately 59 zettabytes in 2020) and use across a variety of industries has led to an increase in demand for data servers and data centers. Six countries house the majority of data centers:
- US (33%)
- UK (5.7%)
- Germany (5.5%)
- China (5.2%)
- Netherlands (3.4%)
- Canada (3.3%)
These are broken down into different types of data center, servicing the needs of enterprises, high growth business, government agencies and increasingly regular businesses that see their data needs soaring:
1. Enterprise data centers
Most multinationals, large enterprises and data-centric firms will have their own data centers. These provide total control over data access, applications and services, and the ability to customize these to meet their current and future needs. Most were built within the business, but increasingly can be built at specialist facilities to reduce the need to modify a building or take advantage of a data-center specialist area (which explains why the Netherlands has so many.)
2. Hyperscale data centers
Built and used by the Googles, Microsofts and Amazons of this world, hyperscale data centers are focused on massive levels of performance and scalability to deliver global services for those companies or their customers. Each will run thousands of servers with high-speed connectivity, and the builders are constantly focused on data center innovation to deliver performance, building their own networks, connections and hardware to keep costs down for users.
New hyperscale data centers will deliver the “metaverse,” moving beyond crunching big data to deliver highly visual representations of business and consumer interactions for billions of people.
3. Managed services data centers
For the rest of the world’s companies who have more modest data needs, managed services data centers fill the gap, providing the computing resources your business needs without worrying about the management of the data center.
As with traditional IT managed services, there are a number of different types of contract and operating agreements, with managed services for storage, servers and varying levels of management of the systems. For any business that has outgrown its own data requirements, these provide quick and easy upgrades to data center-class performance without the need to worry about building one.
4. Colocation data centers
Serving the enterprise and hyperscale market, colocation data centers are built by third parties, often located in clusters in data-center friendly areas with good power and networking facilities, that aren’t prone to natural disasters or regular weather events.
The owner sells the data center to enterprises or even up-and-coming hyperscale users looking to meet their data needs and future requirements, while managing the buildings and some services.
The benefits of colocation include removing the need for a lot of technical know-how from the business. They also offer ease of integration with existing services and buyers can rent from a single data center cabinet all the way up to 100s of them, depending on their need.
5. Cloud data centers
Cloud data centers are designed to supply the cloud to vast numbers of diverse customers, often across multiple regions or globally. Therefore they’re more focused on fault-tolerant and redundant networking than on raw power and performance, providing all those cloud services most of us rely on.
Data is shared across multiple servers for redundancy, all working invisibly to users in the background, delivering the cloud as they check their emails, build websites or applications or collaborate over documents. Depending on your business needs, choosing cloud over a data center may have costly implications as the company grows.
What is the future of the data center?
Data centers are far from set in stone. Like any business technology, they continue to mature and develop to better address the needs of users and enterprises.
Data centers have moved to offer greater virtualization and converged (and now hyper-converged) infrastructure. These provide all a business needs including the applications and services required for the company to operate, rather than relying on different providers.
From a business perspective, this simplifies purchasing and management, guarantees compatibility and reduces cost. Adopting converged infrastructure means if a company is growing headcount, you can add more worker resources faster.
Hyperconverged infrastructure offers virtualized compute and storage resources, which enables businesses to scale faster for compute power and storage as a company’s compute demand grows. This is key for AI applications, those working in intensive data markets as well as banking and high-customer volume applications.
The rise of colocation data centers also means fewer companies will choose to build their own data centers, reducing the risk. This is a key factor as due to the growth in remote working, many firms are downsizing offices, and they don’t want to be moving their data center from one HQ or regional office to another.
That said, the data center is still very relevant as more businesses become reliant on massive volumes of data. As we move to the 5G and AI era, business and production automation produce ever-more data.
Firms will be reliant on data centers to process information in real-time to capture insights and alert the business to predicted failures and other deviations. Yet, while the data center becomes more vital, they’re ever-more automated, requiring less oversight, enabling greater remote management as patches and updates flow automatically and user policies are automatically enforced.
All of these advances mean that acquiring Data Center-as-a-Service (DCaaS) will be as easy for businesses as getting up and running with a new cloud service. And as greater commoditization increases, data centers will become cheaper to rent or acquire, enabling larger enterprises to use hybrid data centers, just as they do cloud ones.
How DCIM can help data center administrators
At the heart of any data center is a growing range of data center infrastructure management (DCIM) tools. Using monitoring software, sensors and the health-monitoring features within most data center components, they help administrators keep the data center operating at peak performance, minimize downtime and predict future failure points.
Using data center monitoring with AI built-in further automates these efforts, especially as data centers become hybrid and move to the edge of networks to improve business performance. AI can alert managers to issues faster than traditional systems and help with future planning.
As the need for data centers grows, managers can use capacity planning to schedule future acquisitions or upgrades, modeling the storage and compute requirements along with space for new racks, ancillary equipment or the need for a new data center.
The combination of smarter hardware, monitoring and management services means that data centers are more reliable than ever. They’re easier and more affordable for businesses to adopt, with less IT knowledge required and fewer management overheads.
As firms become reliant on data, the data center – either in on-premises or a colocated environment – will be an essential tool for those looking beyond the simple cloud to manage the load, ensure efficient business operations and provide a path to future growth in the era of very-big-data.