01.SEA-NGThe Underwater Homelab
Welcome to the beginning of an ambitious journey that brings the world of servers into a new, unexpected frontier - the ocean. This exploration is dubbed the “Underwater Server Project,” also known as “SEA-NG”. My mission? To challenge the conventional boundaries of data storage and processing. By merging the power of cutting-edge technology with the wild, unpredictable nature of the sea, I have a unique challenge on my hands.
At the heart of SEA-NG is a simple yet daunting question: Can I establish a reliable, efficient, and sustainable homelab at the bottom of the sea? It’s an inquiry that pushes us to reimagine the potential of server technology, encouraging us to think outside of the box when it comes to data storage and processing.
While SEA-NG is an original venture, I owe a great deal of inspiration to Microsoft’s Project Natick. This pioneering initiative sought to investigate the feasibility of sub-sea data centers, and their findings have paved the way for projects like ours. Microsoft demonstrated that underwater data centers could be more reliable, energy-efficient, and easier to deploy in various sizes near coastal areas that need them. Their underwater data center had just one-eighth the failure rate of a land-based data center, a significant improvement that highlights the potential advantages of underwater servers.
Now, it’s my turn to explore what’s possible. This project is in its infancy, and there’s plenty of research and planning on the horizon. From selecting suitable hardware, designing the server housing, choosing an efficient cooling system, and considering the multitude of environmental factors such as saltwater corrosion, deep-sea pressure and temperature, and logistical challenges of deployment and maintenance, I have my work cut out for myself.
A central component in my project is the pressure cooker pot, which will serve as the housing for my server setup. Strong, durable, and designed to withstand pressure, it forms an integral part of my underwater server. You can view the exact model I’m using here .
The SEA-NG project is a community endeavor, and I’ve been fortunate to receive a wealth of feedback and suggestions from the community. This feedback is helping to shape the project, underscoring the power of collective knowledge and collaborative problem-solving.
As I set sail on this journey, I’m acutely aware that I’m venturing into largely uncharted waters. There’s much I don’t know, and I’m certain to encounter challenges and obstacles along the way. But, just as the sea is full of mysteries and untapped potential, so too is this project. I’m excited about its potential, committed to learning and experimenting, and ready to push the boundaries of what’s possible.
Thank you for joining me on this journey. I’m eager to see where it leads and what discoveries await us in the depths of the ocean.
What are we trying to achieve and why? Aside from exploration and challenge, there are some known advantages to this approach based on previous research and experimentation, particularly Microsoft’s Project Natick:
Energy Efficiency: Underwater data centers can leverage seawater for cooling, which can be more energy-efficient than traditional air conditioning systems used in land-based data centers. This is important because cooling is one of the major energy expenditures in a data center.
Sustainability: By using the ocean for cooling, underwater servers can potentially reduce water usage. Traditional data centers use a significant amount of water for cooling.
Reduced Latency: By placing data centers underwater near coastal cities, data has less distance to travel, leading to faster delivery of services.
Reliability: Microsoft’s Project Natick found that the failure rate of their underwater data center was just one-eighth that of a land-based data center. This is likely due to the controlled environment, which lacks issues like corrosion from oxygen and humidity or large temperature shifts common in land-based data centers 1 .
AREAS OF RESEARCH
There are several crucial areas of research that have been identified as important to the success of this project. These areas cover a broad range of topics from cooling solutions to server configurations.
1. Cooling Solutions
Effective cooling is a critical part of this project due to the intense heat generated by the server components. The underwater environment poses unique challenges and opportunities for cooling, and we’re exploring various innovative solutions.
- Passive Cooling: One idea is to weld and expose radiator fins to sea water for passive cooling. However, the feasibility of this approach is under investigation because of potential challenges.
- Non-Corrosive Thermal Agents: Another proposed solution involves using non-corrosive thermal agents like 3M’s submerged cooling stuff or mineral oil. These substances are capable of absorbing heat from the server components and transferring it to the surrounding water through the pot, effectively turning the entire enclosure into a large heat sink. Though I worry about environmental concerns if there were to be a leak.
- External Cooling Mechanisms: An alternative approach is to use an external cooling mechanism such as a radiator, or a coil of pipe making a sleeve of copper tube wrapped around the outside of the pot. This method relies on the high thermal conductivity of copper to facilitate heat transfer from the server to the surrounding water.
2. Server Configuration
Setting up an efficient server configuration that can operate reliably in an underwater environment is another critical aspect of this project. Various options are under consideration.
- Micro-PCs: One proposal is to use 2-3 micro PCs capable of running software like XCP-NG or Kubernetes. This setup would be compact and efficient, reducing the amount of heat generated and therefore the cooling requirements.
- Raspberry Pi Cluster: A more unconventional idea is to create a Kubernetes cluster using multiple Raspberry Pis. This setup could provide a flexible and scalable solution that is also compact and energy-efficient (but whats the fun in that).
3. Protection Against Water Ingress
One of the most significant challenges of an underwater server is protecting it from water ingress. This problem requires careful consideration of cable entry points and ensuring that power and data cables do not pass through the housing.
Potential solutions include using resin potting and slip ring connectors. Resin potting provides a durable seal that prevents water from entering the enclosure.
4. Name of the Project
Choosing a suitable name for the project is more than just a branding exercise. The name should reflect the nature and objectives of the project, while also being catchy and memorable. Two names have been proposed so far: SEA-NG and Project Atlantis. Both options are currently under consideration.
5. Examination of Similar Designs
Even though this project is unique, there is value in examining similar designs for insights and inspiration. For instance, the designs of sensor/data buoys could provide useful ideas, even though they operate on the surface rather than underwater.
6. Safe Materials for Environmental Protection
An important aspect of this project is its potential impact on the marine environment. We must ensure that any fluid used inside the system is safe for marine life, in case of hull breach.
Part 1: Design & Implementation
The first phase of our project - the research phase - was all about gathering information, exploring possibilities, and identifying potential challenges. Now, in the design and implementation phase, we’re focusing on turning all that theoretical knowledge into tangible action. This is where we bring our Underwater Server Project, SEA-NG, to life.
Designing the Server Setup
Our first task in this phase is to design the server setup that will be housed in our chosen pressure cooker pot. We need to ensure that the setup is compact, efficient, and suitable for the challenging underwater environment. The consensus is to use 2-3 micro PCs. These small form factor PCs will be mounted in a triangular frame within the pot, allowing for efficient use of space and easy daisy-chaining of the cooling lines. To support various applications, we’re considering running XCP-NG or Kubernetes on these micro PCs.
One of the critical challenges in this project is cooling. Traditionally, servers generate a significant amount of heat that needs to be effectively managed to prevent overheating and subsequent system failure. Underwater, this challenge takes on a new dimension. We’re exploring two main avenues for our cooling system:
- External Radiator Fins: By exposing the radiator fins to the seawater, we could leverage the natural cooling properties of the water. However, we need to take into account the corrosive effects of saltwater on the radiator. A possible solution might be to use a custom external coil of pipe that can act as a radiator without exposing sensitive components to the seawater.
- Submerged Cooling: A more ambitious solution involves filling the pot with a non-corrosive thermal agent such as 3M Novec or mineral oil. This would effectively turn the entire enclosure into a heat sink, ensuring excellent heat dissipation. However, sourcing these materials and ensuring they work well with our setup will require further research and testing.
Cabling and Connectors
Cabling is another crucial aspect of our project. We need to use armored cabling or conduit to protect against abrasion and potential severing of the cables. The ingress points of these cables are a significant risk of water ingress, so we’re planning on using resin-potted bulkhead connectors at the dry/wet interface. These connectors provide a waterproof seal to ensure there are no leaks.
Once the design is finalized, we’ll proceed with the implementation. This will involve assembling the server setup, installing the necessary software, and setting up the cooling and cabling systems. At each step, we’ll be testing thoroughly to ensure everything is working as expected.
Testing and Deployment
The final step in this phase will be testing and deployment. We’ll conduct rigorous testing to ensure our setup can withstand the underwater environment and operate efficiently. Once we’re confident in our setup, we’ll proceed with deployment.
This phase is where our project truly comes to life. It’s an exciting time, filled with challenges and opportunities. We’ll be learning a lot along the way, and we’re looking forward to sharing our progress with you. Stay tuned for updates as we continue our journey into the depths of the ocean.
Stay tuned to follow the progress!
Last Updated: October 12, 2023