The global market for big data and analytics is valued at $274 billion. Every day, approximately 2.5 quintillion bytes of data are generated worldwide, resulting in over 44 zettabytes of data in the entire digital universe to date. 70% of this data is generated by users.
End-user expenditures on cloud computing amount to around $500 billion annually.
Statistics from leading analytical companies unequivocally demonstrate that big data not only forms the basis for the development of future technologies such as artificial intelligence but also serves as a catalyst for the digital transformation of businesses.
But what exactly are big data, and why are they expected to play a crucial role in the evolutionary development of a large part of industries? What challenges do big data pose to companies before they become a source of competitive advantage? Where do big data reside, and what are the requirements for their storage and processing? Andrey Rassiysky provides answers to all these questions.
Andrey Rassiysky holds dual specializations in digital strategy and innovations from the Massachusetts Institute of Technology (MIT) and master's degrees in business administration from IE Business School and computer science from New Bulgarian University (NBU).
In recent years, one of the main technological terms is "big data." What lies behind this term?
This term describes structured, semi-structured, or unstructured information generated by the operations of various industries, with varying value in terms of how businesses are conducted. There are companies for whom data is crucial, while others generate data but may not use them as extensively.
The standard definition of "big data" includes the so-called 3Vs – volume, variety, and velocity. Volume refers to the amount of data, variety to the diversity of data types, and velocity to the speed at which data are generated, stored, and processed. Data themselves also have the property of generating new data. Once generated, certain data reproduce, leading to an exponential accumulation process.
Over the past few years, big data is omnipresent. They are generated by almost every human activity and can be both a challenge in terms of storage needs and an advantage when analyzed.
What are the trends in this technological field?
Big data is associated with a phenomenon known as the digital economy or the economy of bits, which emerges based on the combination of three technological pillars – machines, platforms, and distributed knowledge.
In this context, the term "machines" is used as a comprehensive concept for various modern technologies, including artificial intelligence, 3D printing, drones, neural networks, etc. The second pillar – platforms – encompasses all businesses utilizing platform business models, such as Uber.
The idea behind these types of platforms is to bring together supply and demand in one place and provide them as a service. In the case of Uber, the company connects people seeking transportation with those wanting to work as taxi drivers.
Another example of a similar business model is Airbnb. Many individuals temporarily or long-term rent out apartments, while others are looking to rent. They connect through the platform, with one group receiving services from the other. As a result, Airbnb has a larger market capitalization than any current global hotel chain, despite not owning a single hotel.
This phenomenon is the essence of the digital economy – platform business models enable the creation of businesses that do not own assets in the traditional sense but are so powerful that they surpass traditional sector representatives.
Currently, 5 out of the top ten companies by market capitalization are based on platform business models (a far cry from the situation 20 years ago). They are exceptionally influential, transforming the entire economy and the way business is conducted.
As for distributed knowledge, it refers to a process that does not rely on the expertise of a single company but gathers small amounts of information from millions of people in a given country, region, or location to draw the complete picture. The concept is similar to torrents and is called crowdsourcing.
These directions form the basis of the digital economy, but each has an analog in the classical economy. Machines find their analog in human intelligence, platform business models in the products and goods people produce, and distributed knowledge in centralized knowledge, as each company strives to create such knowledge about how it conducts business.
The digital and classical economies are not in conflict; they complement and evolve together, working towards advancing humanity. Big data serves as the nourishing environment for the digital economy. For ChatGPT, for example, big data are analogous to what surrounds a small child in the world.
In other words, there is no digital economy without big data.
How do all these big data help businesses?
If modern businesses want to participate in the digital economy, they must inevitably harness big data. Without the ability to perceive, analyze, process, and generate large amounts of data through one of these three groups of technologies, they cannot engage in it.
The classical economy or the economy of atoms emerged in parallel with the rise of human civilization. However, the economy of bits began with the advent of computers and the internet, marking the onset of big data generation. At its core lies the so-called digital commodity, without which there can be no digital economy.
Every commodity in this group must meet three basic characteristics – it must be perfect (each copy must be identical to the others), it must be able to spread instantly, and the costs of reproducing it must be close to zero.
Classic examples of digital commodities include software products and MP3 files, while analog products do not meet these requirements.
If a business wants to enter the digital economy, it must use data as a nourishing environment and define the digital commodity that forms the basis of its business model. It is precisely data that helps form this definition.
Which are the most data-intensive industries, and what necessities does this intensity create for them?
Industries with intensive information and data usage are those that generate massive amounts of data, use them intensively, and which data loss would be critical for from a business perspective. The analysis conducted at "Telepoint," based on the OECD classification, identifies among the leaders the entire sector related to ICT—telecommunications, software companies, etc.
The financial sector is also a highly data-intensive industry due to the generation of enormous data volumes.
Another example is healthcare, where regulatory requirements mandate healthcare institutions to retain collected information for a specific period of time.
All marketing research, business services, and advertising also fall into the category of industries with a high intensity of data usage.
Examples of the opposite are agriculture and the mining industry, where data are generated but not on such a scale.
Regarding Bulgaria, according to our analyses, approximately 24% of the Bulgarian economy is characterized by intensive use of big data. This includes telecommunications and ICT, machinery and equipment manufacturing, wholesale and retail trade, repair activities, financial services, healthcare, and more.
Colocation and cloud services are fundamental solutions to the challenges that big data pose to businesses. What advantages do they offer over local data centers?
Data centers are the places where big data reside. While data are generated on the internet, they need to be stored somewhere, and businesses have various approaches to this. The most generic approach is to build a local data center, complete with the necessary facilities such as an uninterrupted power supply, diesel generators, air conditioning systems, fire suppression systems, security, and so on.
Another approach is to turn to providers like us, who create installations housing hundreds of thousands of servers and server racks.
The difference comes from the economy of scale. When we do it, we do it for a large quantity of equipment, while a bank, for example, might do it for 30-40 server racks (the quantity needed by financial institutions in our country). The cost difference for one server rack between us and them is over three times – ours is three times lower than theirs.
Furthermore, after the initial investment, there are ongoing maintenance costs. Our experience suggests that any project below 100 server racks as an investment is not profitable. From this perspective, any business deciding to do things on its own should know that projects below this scale are unprofitable, and it is much better to use the services of an external provider like us because they will receive all these systems at a highly professional level with guaranteed support.
Another example is mobile operators. The software for a mobile network requires around 30 server racks, i.e., not many companies need 100 server racks. This is the essence for the existence of colocation centers – gathering everything in one place, providing highly professional services, maintenance, and systems. Users get something that would be challenging to achieve on their own in terms of quality, and from an economic perspective, they would never achieve profitability on such an investment.
What are the requirements for a provider of such services? What must they have to be fully useful for their clients?
According to the classification of colocation centers, there are six groups, with the last two (5 and 6) exclusively includes the centers of technological giants like Meta and Google. Mass-distributed centers are divided into four categories, and the categorization depends on factors such as guaranteed system uptime.
When a company has servers and maintains them itself without special requirements, it is generally considered an industrial power user and draws power from a single substation. However, in the event of a power outage, to keep the servers running, they need an Uninterruptible Power Supply (UPS) to take over power supply until the diesel generator kicks in.
To avoid this scenario, a connection to a second substation is necessary, as required in facilities like hospitals.
Service availability is measured across several categories, with the lowest being 99.7% on an annual basis. In Category 4 data centers, the indicator is 99.995%, meaning less than 26 minutes of downtime per year on average.
In Category 1, the downtime exceeds 28 hours per year, a significant difference.
This example extends to other systems like air conditioning.
The higher the category, the higher the level of services provided, but also the investment required. As already clarified, profitability is achievable only with a sufficiently large volume. Since most businesses cannot ensure this, they start compromising on quality and category.
Therefore, instead of, for example, a bank having a Category 4 colocation center, it might settle for a Category 2 local data center, risking over 20 hours of service interruption annually. Even if the center is Category 4, the investment is substantial and unprofitable.
How is the colocation services market developing in Bulgaria? Has the Bulgarian business community recognized their value?
To a large extent – yes, but it also depends on the specific industry. ICT companies quickly grasped the concept because they themselves heavily utilize data. Healthcare, financial services, and even automotive parts manufacturers are also beginning to understand the benefits of colocation services.
The market for colocation services alone is in the range of 30 million euros in annual revenue, but when including various cloud services, it exceeds 600 million euros. For comparison, the market for telecommunications services is over 1.5 billion euros.
In short, the market still has a lot of room for development. The fact that the sector is evolving indicates there will be growth. So far, we've seen an annual growth rate of around 17%, but since its establishment, "Telepoint," for example, has registered double the average market growth rates.
In Bulgaria, there are 12 companies offering similar services, with 22 data centers built across various categories.
As this business is not regulated, companies are not obligated to provide data to the authorities, and market share estimates are made indirectly based on declared data from providers regarding colocation space and the amount of provided power. The statistics place "Telepoint" in the top position, which is understandable given that the company is one of the pioneers in the market, holding approximately a third of the market share.
The future of big data and, specifically, Telepoint:
The first machine era, also known as the Industrial Revolution, amplified human efforts both physically and, with the advent of computers and the internet, mentally. Computers multiplied human cognitive efforts, significantly accelerating the thinking process. This led to the emergence of big data, generating more information in a few hours on the internet than in an entire human lifetime during Shakespeare's era.
As we enter the second machine era, businesses now have a choice: to remain outdated or to embrace the byte economy, enhancing their competitiveness. The rise of this digital economy is transforming the entire functioning of human society and civilization. Those unable to adapt risk fading away due to pure evolutionary reasons.
Big data not only serves as the nourishing ground for the digital economy but also transforms the way business is conducted. Managers can no longer be confident in their decisions unless based on big data.
As for Telepoint, without companies like us, big data in such volumes might not have been possible. We continue to invest in colocation centers but also look toward the future.
There are trends in data storage and processing, including centralized locations and edge computing, where data is processed and stored at thousands of endpoints. We are in talks with several companies servicing such infrastructure.
Another development direction for us is expanding our cloud services offering. Cloud services are popular among businesses because, instead of investing in servers and disks, they purchase a cloud service that offers virtualized servers, virtualized disks, and more.
There is also an increase in demand for hybrid cloud services, combining leading cloud platforms such as those from Microsoft and Amazon with a private virtual environment. We install the corresponding equipment for the client, virtualize it, and they use the virtual environment instead of buying physical machines.
In other words, colocation serves as the foundational infrastructure upon which we strive to build additional high-value services.
As of today, Telepoint operates three data centers across the country—two in Sofia and one backup center in Montana. They are connected via a high-speed DWDM network with a capacity of 96 x 100 gigabits for point-to-point data transfer.
Telepoint has formed a group of partner companies around its core business of colocation and cloud services. The idea is that, beyond offering colocation and cloud services, other organizations complement the portfolio. When a person brings their equipment to the colocation center, they can access additional services such as connectivity to the internet and other providers, protection against cyberattacks, and more. Offering colocation services alone is not enough. We strive to cover the entire ecosystem.