As companies scale, they typically migrate from a monolith to a microservices architecture. Because there are no better options available, these migrations often require a company to repurpose three of their most senior engineers on a six-month-plus endeavor to build the underlying infrastructure platform that will support the microservices. Building this platform isn’t easy or cheap, and it can be highly disruptive. Maintaining it is sometimes even worse.
Nucleus is a platform that enables developers to automate building, deploying, and managing microservices in their own cloud environment.
There are many challenges that companies face as they scale from Seed → Series A → Series B and beyond. One of the most potentially disruptive and complex challenges is figuring out how to scale an engineering organization’s productivity as the company quickly grows in engineering headcount and codebase size. Typically the monolithic codebase that many companies start with doesn’t scale well because it’s tightly coupled together. This means that developers have to coordinate changes and releases, which slows them down.
One way to solve this problem is to split up the monolithic codebase into smaller services (i.e, microservices) that allow teams to work much independently and much faster.
Figure 1. The monolith on the left is tightly coupled, while the microservices architecture on the right is much more independent.
In fact, 64% of companies, such as Netflix, Stripe, Uber, Linkedin, Box and others, use microservices. And they’ve seen pretty amazing results. They can ship code 2x as fast and those applications are 29% faster than monolithic applications. More importantly, those applications increase customer satisfaction by 30% because they’re faster and more resilient to bugs and crashes.
Using microservices isn’t without its challenges. Because there are many more moving pieces now, developers and DevOps engineers have much more complexity to manage. In order to manage this complexity, companies need a microservices platform.
Taming Complexity with a Microservices Platform
Microservice platforms help developers quickly build, deploy and manage microservices without having to interact with lower-level infrastructure components such as pods, instances, or networking configurations. This helps developers ship services much faster.
Figure 2. The typical components of a microservices platform.
The problem is that companies have to build this in-house because there isn’t a better option. On average, it takes 2-3 senior engineers 6+ months to build a microservices platform and then another 1-2 DevOps engineers to maintain it. That’s an upfront cost of $400k and an ongoing cost of $200k-$300k. Companies are allocating their best resources to building and maintaining infrastructure instead of shipping features. Additionally, there are major security vulnerability consequences if companies get it wrong.
Nucleus will eliminate the headache, the cost, and the friction of adopting microservices, so developers can build more and configure less. This is
Nucleus: Automating Microservices Infrastructure for Developers
We started Nucleus after facing this problem firsthand at our last companies and wanted to take our learnings to build a world-class platform that other companies can use to scale without limits.
Figure 3. Nucleus architecture diagram
Nucleus is an integrated product that sits on top of a customer’s infrastructure and gives developers a way to build secure, enterprise-ready microservices in minutes. We solve four main problems:
Helping developers have end-to-end observability over their services
Securing services with public and private authentication (tls + mTLS)
Protecting services with a least privileges, authorization framework (think OPA, OSO)
Providing a gitOps-safe way of developing services and protecting secrets
Developer Experience = Simplicity + Control
We deploy the Nucleus Data Plane into the customer’s account while we manage the Nucleus Management Dashboard. Developers can easily configure and use Nucleus using the CLI and Dashboard. When they want more control and visibility, they have access to the lower-level platform components in their account.
Figure 4. Nucleus deployment model gives developers a better experience without losing control
This has three major benefits:
We reduce complexity where developers want it but give them control where they need it. Developers can use the management dashboard to configure and manage their account if they want the “easy button”. Alternatively, because the data plane (where the microservices run) is in their account, they also have access to the individual Nucleus components if they want more control and deeper access.
The customer takes care of the infrastructure bill. Infrastructure is easily the most expensive part of a company’s cloud bill, and because we deploy the data plane in the customer’s account, they pay for the infrastructure. This means Nucleus’ COGS to deliver the platform are quite low. This will help us get to profitability faster.
Very high switching costs. Once we’re deployed within a customer’s environment, and they’re running mission-critical microservices on the Nucleus platform, it will be very hard to rip out Nucleus. That would mean risking downtime and an outage.
Focusing our GTM on Mid-Market
The mid-market is the right place to start for two reasons. First, they feel the need to get to market faster and increase developer productivity most acutely. They’re typically growing quickly and re-architecting for growth which makes it a natural place to consider adopting the Nucleus platform. Secondly, they have a budget allocated for infrastructure products and tend to make decisions much faster than larger enterprises. This is a strong market that is growing quickly.
Led Platform team @NewFront Insurance as Staff Engineer
Designed and built an internal microservices platform
MS CS @Georgia Tech
We want to own the end-to-end microservices lifecycle
Long term, our vision is to own the end-to-end microservices lifecycle. We’re starting with Build and Deploy and then will expand to Test and Manage over time.
Additionally, what is powerful about our approach is that once developers deploy services onto Nucleus, we have the atomic-level unit (the service) from which other services and products are derived from. One great example is the service catalogue. Once we have services on Nucleus, it’s easy to create a service catalogue that helps team understand what services they have available and who is responsible for them. This means that we can effectively compete with companies such as
by expanding horizontally. This presents a huge opportunity in an adjacent market that we will be well-positioned to win.
Why will we win
We already have experience building these platforms. We’re a strong product and engineering team that knows what it takes to build this platform well.
From a competitive standpoint, others in the space only offer partial solutions, and the ones that are more complete take months to implement. Nucleus is a complete solution that takes minutes to get started.
Lastly, and maybe most importantly, we’re scrappy, competitive and want to win. We have a chip on our shoulders and want to prove to the world that we can do this. We want to build a product that others love and something that we can be proud of. We want to build the next billion-dollar company.
We’ve also put together a quick demo showcasing a few features. Check it out below!