Conversations with Joshua Au

While my experience may not speak for the rest of the industry, day to day, I am involved in mainly administrative work, where I am in charge of ensuring the vendors are doing their work. These vendors come from various categories, where there could be mechanical engineers, electrical engineers, and service-related engineers. It is similar to how hotel management would run their hotel establishments, where the service operator is in charge of all aspects revolving around the customer. In our case, the customers purchase the storage and servers at our data centres for their IT and network application purposes.

The main challenge is that there is no straightforward or simple approach to dealing with vendors and handling operations. As requirements, pace and complexity evolve, managing the ecosystem of varying vendors, working style and structural differences in data centres makes the job dynamic. The work involved can differ depending on the data centre one is working at, despite having similar goals of providing securely backed-up servers and network storage.

The short answer is no. At the point of my graduation, there was no such thing as a commercial data centre. The job was created due to the needs of the industry and the growth of the internet economy. I went back to trace contacts of people who eventually ended up working in this particular industry. Through my research, I found that it was first started in Europe, mainly to focus on industries like finance rather than exclusively on the Internet.

Bearing in mind that back then, Philippines, Indonesia, Singapore were equally new to the data centres. As the industry grew in the region, most recognized the need to manage more factors for servers to work, such as electrical distribution, and the ambient temperature and relative humidity of the server room though the telecom sector had a head start given the first data centres were primarily an extension of the telecoms industry. As a result, many people who joined have a background in Tele-Communication data centres or pre-sales for IT departments. Especially in the last 5 years, the growth of the Internet has shown that data centres are here to stay, as more people are going into a career in data centres. However, in the early years, we mostly ended up inadvertently in the job. Personally, I came from a background related to IT services, dealing with backups, writing scripts for servers and I wanted to move to something where I could clearly see the backend operations of the internet.

Initially, after my graduation, I spent a year at Sun Microsystems in the customer service department. I then went to a company based in Hong Kong where we had a chance to travel to places like Indonesia, The Philippines and China. There were opportunities to get deployed to data centres, where I found out that these centres were hyper-localized and differed based on the regions and related regulations. There are various aspects to the job, so I had to familiarize myself in electrical signaling, cybersecurity, project management and operation management. The combination of deep training in a variety of related areas gave me the technical expertise to manage data centres. That technical expertise, which qualified me for the job, coupled with my aptitude for handling things meticulously, resulted in me developing a passion for this job.

To give some background, I was working in data centres not only in Singapore, but also in the Philippines and Indonesia. I could tell that some of my friends in other parts of the world knew that data centres would be a hit business, especially people from Europe as they were early adopters. For the rest of us, we had to slowly figure out what it meant to operate data centres and centralize all of the operations. The initial server rooms only held telecom equipment. It then slowly scaled to include tools to support the data network. These included remote storage and servers that were added to the system. However, in order to allow this incremental increase of equipment to be run in a reliable manner, the requirements were above and beyond what most companies or users were familiar with. For example, in a normal shopping centre, we are only concerned about air conditioning if it is sufficient at a wide range of temperatures. However, in the case of a data centre, there is a need to conform to certain recommendations for temperature range and prevent the relative humidity from varying too much because it may damage the storage tapes, and cause other complications. In addition, there could be load management of aspects like electrical distribution. Thus it is a lot more complex, and most efficient to centralize the operations into a data centre rather than in individually managed server rooms. Slowly but surely, I realized that there were going to be hundreds of thousands of people doing the same. The term data centre now is extremely loaded. It is now a complex concept, where no one person can claim to understand it all. That complexity also means that many industries are now dependent upon it and it certainly has a big role in today’s economy.

Furthermore, with the explosive growth of the internet, the demand for data centres, securely managed and backed up has exploded. It is now clear that the data centre is here to stay. A data centre is only possible if there is the right mix of people running the data centre. That was how I knew that this was an industry here to stay and requiring skilled manpower to support it.

The industry in general is eager to hire ex-army veterans. Army veterans tend to have the mindset of professional skepticism (mission-critical paranoia) or minimising risk. Our job is one where even the slightest risk taken could jeopardise millions of companies. Risk-averse profiles are probably the best fit candidates for the job. They tend to make less assumptions, and then as the process propagates down, that particular nature ensures that the process runs well. For example, if a data centre is responsible for supporting a company that runs a live streaming service and due to negligence, the service goes down for 10 hours. This could mean the end of the company as many of the viewers would likely become frustrated with the disruption and cancel subscriptions. After 9/11, many companies went bankrupt because they lost their IT and data infrastructure after the plane crash into the World Trade Centre.

Today if you are a bank, hospital or any mission-critical infrastructure service, there must be some element of IT services in terms of managing records, managing inventory or managing data. It would be unthinkable to imagine that those companies or organisations would stop working as a result of server stoppage.

I eventually realised that the people who join this industry need to prepare to stay in this career for a long time to build up the required expertise. This is the case not only for data centre operations, but for IT services and technology in general. For example, today, people can learn a specific framework for coding. 10 years from now, the same framework may not even be relevant due to the rapid pace of innovation within the industry. Precisely because the development ecosystem is changing very fast, it is very hard to keep pace without spending a long time learning the exact principles and continuously refreshing your knowledge. However, whatever actions the organization wishes to execute, it is only possible if the servers are ready. The people who maintain the servers need to be capable of responding to the changes in the development ecosystem and continuously evolve the servers. This is crucial to support changes in the ecosystem and new languages that grow in prominence every year.

The short answer is that it is not possible to fully dominate the data centre market. It is not easy even for the large internet giants like the FAANG (Facebook, Amazon, Apple, Netflix and Google) companies. Although they bring massive demand for server usage, establishing a data centre is not a process that is replicable for every location.

Most of the Internet giants own their data centres, but in some cases they prefer to co-locate. For certain business cases, there is no good reason to establish your own data centre. It is not as efficient and results in a slower go-to-market time gap. The expedient thing to do is go directly to local vendors. For instance, there are differing local codes of compliance to follow in different countries even for the same exact usage scenario.

That kind of regulatory knowledge is something so specialized that very few actually understand how to deal with them quickly and efficiently. For instance, to build a data centre in India, you need to obtain more than 40 different permits. That translates into a possible need to talk to more than 40 different agencies and departments to get the right permits. In Singapore, it numbers in the 20s. For corporations, the regulatory framework slows processes down and increases the risks of not being able to run the business, which results in a specialized need for localized data centre operations and management (which knows how to navigate under these requirements) to exist.

Simply put, cloud and edge are about meeting customer expectations. As customer expectations have changed, the possible use cases have shifted as well. Getting information used to be an extremely laborious task, but now with the Internet, the demand for information is largely instantaneous. Similarly, the primary reason for edge computing is that bringing the data centres to the average consumer greatly speeds up the latency and the speed of processing. Instead of having to send information to a server that might be halfway across the globe, it is possible to process and compute locally. That translates to convenience and instantaneous response, which is the expectations that consumers have for internet-related services these days.

The convenience is truly addictive. And because of that, it creates very interesting use cases. Devices placed closer to where the service or data is to be consumed make it possible for companies to cache part of the static data locally and operate based on it. Developers generally try to reduce their latency in order to improve the experience. Therefore, this requires the computer to move closer to consumers.

A relatable example could be live streaming, where the latency impacts much of the experience. It does not make sense that for a consumer to be watching something in Singapore but have the server be based in Finland, incurring additional time for the data to reach the consumer. Customers are willing to pay for the experience, and where the money is, the industry will follow. The same applies to telemedicine and industrial IoT (Internet of Things). Edge computing will continue to grow in importance, because that is where the money will go.

Once edge computing is set up, the last mile optimization of speed is on mobile broadband. That is known to most people as 4G today, soon to be 5G. If the mobile broadband speeds are sufficiently fast, it is possible to live stream with mobile phones instead of plugging in cable for guaranteed internet speed. Live streaming would be possible anywhere on the go as opposed to a fixed location. However, in order to do that, it would be necessary for servers to be situated geographically closer to process the data because of the latency involved and that is where edge computing becomes important.

As for cloud computing, the term is rather complex. There are many different scenarios and types of technology involved in the cloud. Ultimately it is about the backend operations done by a server from another place to provide services to the end-users. However, what I would suggest is that where there is a relevant use case and people are willing to spend money on it, that is where the focus of the industry is going to be at moving forward.

People think that data centres do not exist. Digital infrastructure is not frequently talked about because it is not the most interesting aspect to think about the backend processing and operations that the Internet relies on. However, if it does not exist, the Internet will also be impossible.

Typically, if you consider e-commerce, people will still think that there must be a fulfilment order station, inventory and delivery stations etc. In the public domain and schools, we discuss the internet mostly in terms of the network applications, but not the infrastructure. It is important to know that the data centre is the backbone of all your treasured digital services and it is here to stay.

Data centres were already built to be resilient in terms of structure, network, power and people.

The resiliency I speak of is not just physical or in terms of people – we also need the resources. In fact, I know a particular data centre that overcame the limitations of lockdown during the pandemic period by having a staff member spend his lockdown in the facilities. By making it possible for people to physically stay on its premises, this ensures that the data centres can continue operation in the case of a pandemic or natural disasters. This is especially important when the data centre has to support essential services like healthcare. For instance, in Japan, data centres can withstand earthquakes up to a certain reasonable point. As we are trained to assume the worst, we aim to prepare more.

That is why some data centre operators in different parts of the world expect to build self-sufficiency even during emergencies such as typhoons and other natural disasters, since there is such a high level of dependency on data centres for services to continue.

When it comes to resiliency, it is imperative for us to manage these nearly impossible situations. Can it be done? Yes. That is why data centres invest so much in business continuity, risk management, and crisis management. Just one disaster and the company could lose its leaders, manpower, and infrastructure.

It is crucial that those working in data centres realize the impact they have on the organisations their centre is supporting. If you fail them, you may lose your job, but life goes on for you. But for them, will there be a case where their organization closes down? Consider essential services – if a hospital cannot run as per normal, it will affect a lot of lives. Thus, those who are interested in this field of work have to recognize the responsibilities their jobs come with. This attitude of assuming responsibility and remaining vigilant is very important in my line of work.

Aptitude is also important when it comes to learning. A data centre manager may not necessarily need very technical skills, especially in the more managerial positions, but one must know enough to communicate well with the technical employees and earn their respect. You must have enough knowledge to safeguard the interests of the organization while working with external vendors. It is important that you know enough to ask the right questions to keep the vendors on their toes. In fact, I begged my boss for my first job – I knew nothing but I was willing to do whatever it takes because I was and am still passionate about it. To my surprise, he said yes. This happens in the industry. because it is easy to build up that 5% gap in knowledge, but it is very difficult to build up that attitude – the ability to sit up and see, and question critically. What is truly important is to learn the art of asking the right questions to clarify concepts, thinking critically, and the art of managing people’s expectations and these are transferable skills you can bring to any job.

Recently, Karen Tay shared in a webinar that it will be more acceptable to work remotely even for the big companies after COVID-19 (Karen is Regional Vice President (Global Tech Talent) from EDB and based in the San Francisco Bay Area)  In the past, if I wanted to work for the likes of Microsoft or Google, I needed to be based in certain towns where real estate is scarce and rental is expensive. But now because everyone is forced to work from home and they realize it works, IT giants will be more accepting of remote workers. This means that competition for jobs is going to get even tougher, so the challenge will be when we reach a point of feature parity where talent looks the same throughout the world and the only differentiating factor is cost

In this age of competitive global talent, it is crucial that you find your competitive advantage. And if you do not have an advantage in terms of cultivated relationships and insights, geography or circumstances, you need to figure out what can set you apart from the hundreds and thousands of others vying for the same opportunity by using a well thought through game plan. For one, try working with recruiters to build alliances and network with people by giving them value and recommending potential job candidates. Another would be to create a persona both online and in-person, and when you do get the job, live up to the expectations of this persona. Dedicate time to craft your application for each job, and show the companies what they want to know the most, instead of sending the same resume. By focusing your time and resources on a specific advantage, you can help differentiate yourself from the competition.

I think if you have a well thought through game plan, you will be able to work with anybody. So for example, you want to be the CEO of a company, you know that you need to know a thing or two to get started. In the same vein, not everyone is suited for our industry. However, it is important for you to know something about my industry because it will help you to be more shrewd in how you work with your IT partners

If you are interested in joining our industry, feel free to reach out to me with your thoughts, or for advice because I understand what you are going through. Especially if you are not sure about how the industry works, you can find out more about it Because you do not want to depend on luck for something as important as your career.