Data Centers: What, Why and How?

Posted By :
Data Center

In many ways, data centers today stand as a symbol of our data-enmeshed, hyper-connected world. Most people view them as gigantic storehouses of data – but that would be a gross oversimplification.

Data centers form the backbone of thousands of businesses and hundreds of governments – hosting critical operational and proprietary assets in a highly secure, centralised, and increasingly virtualised environment. There are well over 7.2 million data centers in the world as of now. This number does not seem that big when you realise that our planet generates over 2.5 million terabytes of data every day – and that is only going to grow exponentially as 5G, IoT, Industry 4.0, etc. take over.

Given the scale and importance of data centers in our world, it is worth understanding:

  • What is a data center?
  • Why do businesses need data centers?
  • What are the components of data centers?
  • How do data centers operate?
  • What are the various types of data centers?
  • How do data centers contribute to data security?
  • What is the role of cloud in data center operations?
  • What is edge computing and the rise of micro data centers?
  • What are the infrastructural standards of data centers?

This blog answers these questions and many more.

What is a data center?

A data center is typically a physical facility or location that organisations use to house their core business applications, information, and data. Data centers facilitate the centralisation of an enterprise’s IT infrastructure and enable the storage, processing, dissemination, and maintenance of data, applications, workloads, and critical equipment on physical or virtual servers. Consequently, data centers are highly secure and controlled environments that play a vital role in business continuity.

Data centers are in the middle of an evolution that is enabling them to move away from on-premises, physical infrastructure to off-premises, third-party, virtualised edge cloud infrastructure. A combination of the two has made hybrid models of data centers very popular. This tectonic shift in the data center industry is down to the rise of multi-cloud computing across enterprises as well as scalability that suffices current and future data-hungry applications poised to transform our lives digitally.

By 2030, the global data center industry is forecasted to reach a value of US$517.17 billion.

Why do businesses need data centers?

For businesses, data has become the most important asset in a digital economy. As their operations become increasingly digitalised, businesses face a higher degree of challenge in data management, storage, compliance, and security – all of which are vital for organisational agility, scalability, and resiliency.

While a small business can store and manage data such as transactions and emails on a server and even on-premises, larger businesses need specialised data management to support their ever-growing business applications and consumer demands. Doing so on-premises will require a lot of infrastructural investment, maintenance cost, real estate, manpower, special security, and power requirements.

This is where data centers provide businesses with a dedicated, cost-effective, scalable, and secure solution for all their data needs. Providing infrastructure as a service – either physically or virtually – modern-day data centers enable businesses to store, manage and disseminate data for running their applications and services. Additionally, data centers enable the installation of world-class network infrastructure that can support email and file sharing, productivity applications, virtual machines, web apps, databases, etc.

Think of data centers as a business’ resource pool of unlimited operational power. Today, hyperscale cloud data centers sprawling millions of square feet in the area comprise thousands upon thousands of servers to support high-end business demands such as big data, AI and ML, cloud computing, customer relationship management, enterprise resource planning, etc. With such virtualised data centres in play, businesses can share the resources of each server through multi-tenancy while remaining isolated from each other. Ultimately, through such hyperscale data centers, businesses are able to achieve higher compute, processing and storage while using data center resources efficiently and at a lower cost.

Ultimately, data centers take a business’s load of managing enormous data and related infrastructure away, thus enabling them to focus on critical functions, scaling their operations, and improving business outcomes.

Number of Data Centers

Data Centers: Components

Data Center Components

For facilities of such a mammoth scale and complexity, data centers can never be a singular unit. Instead, a data center is a trifecta of three components:

  1. Storage infrastructure

Where are all the enterprise data stored inside a data center? The answer lies in hundreds and hundreds of media such as solid-state drives and tape. But there is not just one copy of each data set. Multiple backups ensure that the stored data is always safe and accessible. Storage infrastructure is responsible for managing all the data required by end-users and businesses for daily operations.

2. Computing resources

The other component of a data center is its computing resources. Think of all the gargantuan processing power, local storage, network connectivity, and memory space required to run enterprise services and applications – all on typically a global scale. This is achieved by high-end server racks.

3. Network infrastructure

The last component of a data center is the network infrastructure that connects it to businesses, end-users, and the fiber backbone. This is achieved through network equipment comprising switches, routers, application delivery controllers, etc.

Providing continuous operational support to the critical components of a data center is a role reserved for the utilities. They include power subsystems, uninterrupted power supply, and backup generator sets. Data centers are also very climate-sensitive as they generate a lot of heat. This means adequate ventilation and cooling systems should always be in place. Lastly, the highest standards of security measures – both physical and software – are reinforced to keep the data and infrastructure safe.

Data Centers: Operations

So, how do modern data centers operate? The answer depends on the type of data center – something we will talk about in the next section. While certain data centers are merely used for storage and data backups, others house the entire core IT stack of an enterprise.

Typically, the operations of a data center are geared towards ensuring application delivery performance and preserving the performance and security of physical components. This is made possible by services like application resiliency, load balancing, automatic failover for the continuous availability of servers, etc. Think of a data center as a hub that connects various environments and serves them with requested data needs – whether it is hosting a website, accessing a file, maintaining workloads, deploying applications, or facilitating services to customers. This is made possible through a system of servers as well as network connectivity equipment.

Types of Data Centers

Modern-day data centers – from minor server rooms to distributed buildings – are playgrounds of the latest networking, computing, and storage technologies. But not all of them are built the same. Over the years, various types of data centers have evolved to become mainstream – each offering specific advantages.

1.Cloud data centers

Cloud data centers are the epitome of the digital era we are living in. Imagine being able to access, store, manage and run all your applications, workloads, and data on virtual servers – without the need for designing, managing, building, staffing, or powering any associated hardware.

A cloud data center is a type of data center that is physically located off-premises or distributed across multiple locations. In such a case, the data center hardware is managed and secured not by the enterprise but by a third-party managed services provider. The enterprise, for its part, gains shared access to virtualised computing resources or dedicated access to certain hardware infrastructure in order to run applications and manage websites and data.

In a way, a cloud data center is essentially infrastructure-as-a-service. One can choose public cloud networks like Amazon Web Services or Google Cloud, or stick to private cloud networks.

2.Colocation data centers

Colocation data centers are rental facilities that allow enterprises to rent fully functional data center space and resources. Naturally, these data centers are not located on the enterprise premises. Thus, businesses can take rack and server space on lease and install them at colocation data centers. All the maintenance, powering, cooling, security, etc., will be managed by the facility itself. Such a facility is ideal for organisations with a wide footprint.

The main difference between colocation and cloud data centers is quite simple. In a colocation data center, the business owns its server inside the rented facility. On the contrary, in a cloud data center, the third party owns all the servers.

 3. Managed services data centers

Managed services data center is a model where a third-party service provider manages, deploys, and maintains data storage, computing, and other advanced services. Such a data center can either be partially managed, thereby giving some administrative control to the business, or wholly managed – a scenario where the third party takes care of everything.

4. Enterprise data centers

An enterprise data center is one that is completely built, deployed, managed, owned, and used by a single entity for its own operations. Such data centers are ideal for organisations that have a highly private facility and unique network requirements, such as big tech companies.

Enterprise data centers can be both on-premise and off-premise. They also comprise many sub-divisions of data centers – each allocated to a different purpose. These include internet, intranet, and extranet data centers.

Data Centers: Security

Security will always be a major talking point when it comes to data centers, given the nature of data and information that they handle. As such, data centers can be targeted by both physical as well as cyberattacks. To prevent the concession of integrity, availability, or privacy of data or intellectual property to malicious entities, data centers, and enterprises are always conscious of applying the most rigid security controls. Additionally, security protocols also protect applications, infrastructure, and users linked to the data center environment – whether physical or cloud-based.

So, what are some of the security systems that data centers adopt at their facility apart from building protection and surveillance? The entire system is based on a zero-trusts analysis methodology. Modern security solutions provided by data centers include IPS, firewalls, data access controls, WAAP systems, WAF, and more.

The Role of Cloud in Data Center operations

According to Gartner, almost 80% of businesses will shut down their on-premises data centers by 2025. Why? The answer can be found in the rise of cloud data centers – an evolution that has been in the making since the early 2000s. Today, more and more enterprises are shifting their workloads to the cloud.

This tectonic shift is evident in the fact that enterprises have started spending more money managing cloud infrastructure services than on physical data centers – something observed for the first time in 2019. The emergence of public cloud service providers such as Microsoft Azure, AWS, and Google Cloud also points toward the popularity of cloud data centers.

So, what makes cloud take centre stage in modern data center operations? The simple fact is that virtual infrastructure offers better economies of scale, lightning-fast speed of deployment, resiliency, and agility. Today, commissioning a virtual data center takes a mere click or two on the internet – something that is not the case when you are building an enterprise data center or going the managed service route. Cloud is also a virtual resource, which means scalability of operations is never an issue. It also provides anywhere access.

Cloud data centers are extremely cost-effective because an enterprise is essentially free from managing, powering, upgrading, or securing its own hardware. Cloud also enables greater flexibility to customise your software and hardware to suit your business requirements. This also promotes greater redundancy.

A good strategy is to adopt a hybrid-cloud approach in which an enterprise splits its workloads amongst public cloud and local data centers – the latter offering a greater degree of control over enterprise data.

The rise of Edge Computing and Micro Data Centers

Edge Data Center

If you think about it, data centers make it possible to process a humongous amount of data and information at breakneck speeds. But the fact of the matter is, that data is continuing to explode. The advent of IoT and digital devices means that new avenues of data are coming to life every other day. For all of these transformative technologies to make sense, the processing of data has to be instantaneous. That means reducing latency as far as possible.

This is where a regular data center might falter because the time taken to send data to and from the cloud means there will always be a lag. Hence, the paradigm of edge computing is on the rise, making it possible to process data closer to the edge of the network, i.e., right where it is being generated and gathered. Edge computing is now entering the data center paradigm as well through micro-data centers that comprise network and compute resources to collect, analyse, store and process data closer to the devices. They are compact units ranging from single racks to large containers and often work on the managed service data center model. The use cases of micro data centers deploying edge computing include 5G deployments, CDNs, IoT, etc.

Infrastructural standards of data centers

Data Center Tiers

Did you know that data centers also have tiers? Yes, there are 4 tiers given by the Telecommunications Industry Association (TIA) to denote the level of infrastructural standards and redundancies that a data center exhibits. This serves as a classification to separate higher performance data centers from more basic ones and acts as a guideline for data center design and implementation.

  • Tier 1: Has an average annual downtime of 28.8 hours. That constitutes an uptime guarantee of 99.671% per year. Does not guarantee built-in redundancies for optimal uptime. Ideal for small businesses.
  • Tier 2: Has an average annual downtime of 22 hours. That constitutes an uptime guarantee of 99.741% per year. Has partial N+1 built-in redundancies by the colocation provider to ensure always-on power and cooling support for operations. Ideal for small manufacturing businesses.
  • Tier 3: Has an average annual downtime of just 1.6 hours. That constitutes an uptime guarantee of 99.982% per year. Utilises a N+1 fault tolerance to ensure immunity from up to 72 hours of a power outage. Thus, it has a higher ceiling of redundancies in power and cooling support compared to Tier 2. Ideal for growing businesses.
  • Tier 4: Has an average annual downtime of merely 26.3 minutes. As for uptime, Tier 4 data remains online 99.995% of the time. Utilises a 2n+1 fault tolerance to ensure up to 96-hour power outage protection. Ideal for government entities and large enterprises.

FAQs

  • Who uses data centers?

Most businesses, enterprises, institutions, utility companies, and government entities that require computing and data storage facilities use data centers. These data centers are either their own or a service provider’s or are based on the public cloud. Simply put, any entity that generates and uses data on a large scale has the need for a data center.

  • What is the difference between a data center and cloud?

While both the data center and the cloud are used to store, manage, organize, and operate data, the former is a physical resource and the latter is a virtual resource. A typical data center is often located on a company’s premises or could be located at another geographical location. In comparison, the cloud data center is off-premises and is managed by a third-party cloud service provider. The cloud can be accessed via the internet. Another area of difference is scalability and costs. While a physical data center requires a lot of space, manpower, and investment to scale up and maintain, a cloud is much more cost-effective in terms of scalability and also offers redundancy as it is onboarded as a service and is taken care of by the third-party.

  • How big is a data center?

One can imagine a data center to be available in various sizes. The largest data center in the world at present is located in China and is spread over a mammoth area of 10.7 million square feet. On the other hand, typical data centers lie in the range of a few hundred square feet.

  • Where are data centers located?

The location of a data center is of prime importance. You may notice that many data centers are located slightly off the city premises. However, the location of any data center is not simply pegged as the outskirts of a city or a town. It depends on a number of factors such as:

  • Availability of continuous, uninterrupted, cheap electricity supply or the possibility of in-situ electricity generation
  • Proximity to business hubs and fiber backbone.
  • Ease of cooling for the data center infrastructure, such as large open space or even underground/underwater area, cooler daytime temperatures and airflow, etc.
  • A highly safe and secure location that is not easily accessible for unauthorised entities
  • A disaster-secure area which does not typically lie in earthquake or flood zones
  • How is a Cloud Data Center Different?

Compared to a regular data center, a cloud data center is a virtual, scalable solution that enables companies to host and access their data and applications on a cloud server through the internet. As such, a cloud data center is a third-party data center and is not located on the company’s premises. It is maintained by cloud service providers such as Google, Amazon, and Microsoft and counts as an Infrastructure-as-a-Service facility. This makes cloud data centers much more cost-effective than an on-premise data center.

  • What is the future of data centers?

The data center services industry is expected to grow to $105.6 billion by 2026, driven by rampant digitisation – especially in industries and manufacturing, remote work, the proliferation of 5G and beyond, and the rise of IoT and AI. In response to the upcoming challenges of hyperconnectivity, data explosion, and latency, data centers of the future will be increasingly converged and edge computing-based. Two other trends we could expect are increased automation and sustainability.

Leave a Reply

Your email address will not be published.