Construction and Operations of Enterprise Data Centers (with Kevin Monahan)

Construction and Operations of Enterprise Data Centers (with Kevin Monahan)

Learn about the evolution of the construction of enterprise data centers in this podcast episode with Kevin Monahan, Director of Regional Operations at West Meta.

The early days of data centers

Kevin’s journey in data centers started in the late 1990s, during the dawn of the internet era. At the time, data centers were not as sophisticated as they are today. The installation of a new device called Juniper, which was a competitor to Cisco, was impressive with its new features such as RJ45 for T1s, T3s, and square things with fiber connected to it. Back then, most of the internet was connected through T1s and T3s, but fiber was gradually becoming more common.

The evolution of data centers

Today, data centers have come a long way since the early days. With the advent of 5G and the Internet of Things (IoT), data centers have become more complex and sophisticated than ever. In this evolving landscape, dedicated servers have emerged as a crucial component. Data centers have become increasingly latency-sensitive, especially in the financial industry. Financial institutions require a high level of precision and speed when it comes to executing trades, and they demand near-instantaneous data transfer.

As a result, data centers have had to adapt to meet the needs of the industry. They have become more advanced, with a focus on faster data transfer speeds, lower latency, and increased capacity. Hyper-scale data centers have also emerged, providing large-scale cloud services and big data analysis.

The importance of network infrastructure

The importance of network infrastructure in modern data centers cannot be overstated. A meter of fiber has been found to equate to 5.4 nanoseconds of latency, making it essential to ensure high performance for applications. This level of precision is especially critical in the financial industry, where even the slightest delay can lead to significant financial consequences.

Image

Water and the design of data centers

There were concerns about water use in data centers due to its heavy environmental impact. Several companies are experimenting with free cooling to decrease water usage, but there is still significant room for improvement in this field.

There is a water crisis in many regions of the country, including New Mexico and Utah, and the need for the industry to take responsibility for its impact on the environment was huge.

Tools and technology in data centers

There is progress made in the area of tools and technology in the data centers, such as better filters to prevent sand from getting into the data center, and the use of free cooling to reduce energy costs. Kevin’s approach to data center management includes a signature sign that reads “Stop, Think, Ask” to encourage workers to prioritize safety and communication.

The importance of progressive thinking

In the data center industry, it is imperative to maintain progressive thinking. Constant adaptation to new challenges and technologies is crucial to remain competitive. Collaboration and communication, both within the industry and with the broader community, are essential to ensure sustainable and responsible data center management practices. Those who fail to adapt to these changes will be at a disadvantage in the industry.

Asset management as an ongoing challenge

Asset management remains an ongoing challenge for many organizations. For example, Kevin’s logistics team relied on paper. After asking the warehouse manager for a full inventory of everything, it took three weeks to get a stack of paper. The paper inventory raised concerns about fat fingers and incorrect information, leading to the adoption of an RFID tracking system that was successful in many sites globally.

SrhK3Wu8qdQvJjw9r0Ugbi3O5g1hTRezogCNRpsBfa 1J AxVK4tcfH4Ah76gqrELHV8TCeL8wyK78r8q3XQMT8NLf3VPbhl4cMeUCJ4tUaXYSQlg3cGs1LDYne9Jd6utECqtKkJa2YQBYss6K9KBck

Image Souce

  1. Asset creation
  2. Registering entering/exiting users
  3. Server to track
  4. Inventory
  5. Search
  6. Booking lifecycle events

The need for cloud-based solutions to control physical assets

Effective physical asset control is a crucial challenge in the current role. One of the challenges is over-ordering and the recycling of equipment that has been left outside for too long. 

To overcome this issue, a cloud-based solution that offers a clear view of the material location is necessary. The solution should be secure and prevent theft, while also providing the ability to track equipment from the warehouse to its final destination for efficient data center asset management.

Standards as a contributing factor to organizational success

Standards are crucial for organizational success, as highlighted by the interviewee. Those who adhere to standards contribute to the success of the organization, while those who ignore them entirely risk failure. It is therefore necessary to implement standards and ensure that they are followed to achieve the desired outcome.

The transition from raised floor to slab in the hyperscale world

The transition from raised floor to slab in the hyperscale world shows that the higher ceiling heights in these buildings allow for downflows to be located in the ceilings rather than under a raised floor. This eliminates the need for raised floors, which can be expensive to install and maintain. Overhead cooling is also considered a more effective option, as it is easier to manage and maintain.

Managing the influx of equipment in the data center

Managing the influx of equipment in the data center is challenging. It’s essential to understand the equipment’s flow and impact on the environment. Formal processes are available for decommissioning management, but tracking the equipment remains challenging.

Equipment refresh and decommissioning

Hyperscalers typically refresh their equipment every five years. While they are not directly involved in the decommissioning process, they do assist in helping hyperscalers remove equipment.

However, what happens to the decommissioned equipment is not something they are privy to. Given the sensitive nature of the data that hyperscalers handle, hopefully, they will shred the drives to prevent data breaches. However, this might not happen due to the cost implications involved.

The focus is often on the infrastructure, with less attention given to the data on the equipment unless they are responsible for it. However, challenges arise when tracking sensitive information on disks, and a solution must be engineered to address this issue.

Data security measures

To address data security concerns, the industry standard for data destruction involves a seven-times wipe, whereby the data is overwritten with zeros and zeros, making it extremely challenging to retrieve any old data. 

For example, Kevin used to switch the drives around in their EMC arrays after wiping them to prevent anyone from putting them back in the correct order to access the data.

Software-defined networking

The use of software-defined networking (SDN) is changing the management of data centers, according to the representative. With SDN, all components are part of a cluster, and the software is capable of resolving issues without human intervention. 

9WkrVfP6ngF PoV2gUfh9BEHvAjabw2muAhBnhnndsBPatHPQOnb13fX3BFYJZy6RJvRyfw2yevzoJRy33S4yDsgmczicLPHYcAVQ45L EQ2QyE9DgIQ1Cqbk2v9tokBunIfzHUdiWhJppqPedOKu4I

For instance, in the case of a failed switch, the software wipes it clean and rebuilds it automatically. If the switch still does not work, the software removes it and generates a service ticket for replacement by a human technician.

Negotiating budgets and workforce during the pandemic

The pandemic has resulted in a downturn in the economy, and data centers have had to navigate these challenges while still growing. There was a lot of back and forth with customers who were paying for services. 

Customers were reluctant to lay off workers, and data centers had to negotiate how they would pay their workers during the pandemic. Data centers had to come up with creative solutions to keep their workers employed while still meeting the needs of their customers.

Labor as the key to data centers

Labor is the most important key in a data center, and data centers cannot afford to lose their workers. It is not easy to find people who are trained to work in data centers. 

Data centers have had to create career paths for their workers, and they have had to find ways to get their workers interested in their work. Data centers have had to come up with creative ways to educate their workers and get them interested in their work.

Creating opportunities for learning

Data centers have had to find ways to get their workers interested in their work. Kevin and his team were creating opportunities for learning for their workers. They were getting their workers involved in financial classes and teaching them about the markets.

 They were also doing a “Know your Customer” routine with their network gurus. They wanted their workers to know what they were plugging in and to have a basic understanding of what the devices were and what they did. By creating opportunities for learning, data centers can keep their workers engaged and interested in their work.

Key takeaways from this podcast

Data centers have come a long way from being rudimentary facilities to becoming state-of-the-art infrastructure. They have become increasingly complex and sophisticated, and have had to adapt to meet the needs of the industry. 

The importance of network infrastructure and progressive thinking cannot be overstated in modern data centers. Furthermore, common tools to manage data center assets remain a challenge for many companies, and there is a need for whole sessions on this topic. 

However, there is also a need to address concerns about the environmental impact of data centers, especially their use of water. Overall, it is clear that data centers will continue to be a fascinating and important part of the modern business landscape.