Creating World Leading Databases with HarperDB SVP Jaxon Repp

Jason Repp is the SVP of HarperDB, a world-leading database and development platform that is leading the charge in terms of performance, flexibility, and ease of use. We had the pleasure of featuring Jaxon on one of our Slogging AMAs.

This Slogging thread by Jack Boreham, Mónica Freitas, Jaxon Repp, Margo McCabe, Terry Jones, Jason Green, John T., Mark Finnigan, Cris Silva, Jack swift and Limarc Ambalina occurred in slogging's official #amas channel, and has been edited for readability.

Hi @channel, it's a pleasure to welcome the SVP of product at HarperDB.

HarperDB is a SQL & NoSQL distributed database and development platform that creates sheer magnitude faster than alternatives.

With HarperDB, you get expectation-shattering flexibility, performance, and ease of use for all deployment scenarios, including cloud, edge, on-prem, peer-to-peer, and multi-deployment use cases.

You can ask him anything about:

Hi! Great to meet you! I want to start with this: 8 STARTUPS?! Wow! Can you tell us a bit about this?

Hi pleasure to have you with us. Can you give us a brief overview of your background?

Hi Jack Boreham & Mónica Freitas!

I started as a developer at a digital agency but ultimately wanted to find a product I could love and nurture beyond the scope of a client engagement. Being able to wear lots of hats, invent an experience, and work with a small team is where I thrive.

Of the 8 startups, I've had two “exits”, 5 “opportunities for learning”, and now HarperDB, which is a whole different level of traction.

I still get to write code sometimes, which keeps me sane.

Tell us more about harperdb!

HarperDB is a distributed application platform with persistence built right in. It combines an extremely performant noSQL data store with SQL semantics and an integrated API server powered by Fastify.

Thank you so why is your database the fastest?

Jack Boreham HarperDB is built on top of LMDB, a lightning-fast, memory-mapped key-value store. We can handle 20k writes/second and 120k reads/second… but the real advantage is that we're built to scale horizontally, so there's no limit to the throughput you can achieve with a cluster… and when you move your application AND data closer to your users, you eliminate the latency associated with a monolithic data store sitting in us-west-2.

What's it like building a startup?

It's a wild ride. From the initial vision to product market fit, being able to invent a new way of doing something, test it in the real world, learn from your mis-guesses, and figure out all the parts of a business beyond the product itself… it's the most fun and challenging thing I've done.

I took a job at a fortune 50 company at one point because my wife was pregnant and I needed health insurance… longest 2 years of my life. Never again.

how would you describe your role at HarperDB?

Mónica Freitas, I'm head of product, marketing, and delivery, so I wear a lot of hats. Ultimately it comes down to staying true to our mission as developers ourselves- make it easier to build and deploy software that scales. Distributed data is a challenging problem, but the rewards when you do it right are better economics for our customers and a better experience for their users. So while our engineering team solves the really hard mechanics of reconciling data across a global network, I use customer feedback and personal experience to figure out how the product itself will achieve all the magic without adding complexity for developers.

Hello! What's the biggest challenge when developing a startup?

John T. The two biggest challenges are building the team and finding product market fit.

I am fortunate to be working with some of the smartest people I know to solve this problem- it's an amazing feeling when you know that your team is fully capable of delivering anything the product needs with a minimum of oversight. There aren't that many of us, after all.

Finding product market fit is the other one. As good as it is in your mind or as a prototype, you have to be resilient when real-world feedback starts telling you some of your guesses were wrong. Startups aren't great for people who can't accept market feedback and pivot… we've had lots of opportunities to learn over the past 5 years, and it's only made the product better… but it can also be a source of frustration if your favorite new feature ends up getting dropped.

Nice to meet you. How do you explain HarperDB in lay terms?

Mark Finnigan, we are a database with a user-configurable API server built right in and the ability to connect as many nodes as you need together to deliver data to end users as quickly as possible, no matter where they are in the world.

Hi, who is your main target audience?

Cris Silva, currently, we are targeting two main customer types: large network providers (Verizon, Akamai, Google, etc.) that can benefit from moving data to the edge to improve their customers’ applications’ performance. Moving data to the edge reduces latency, so it's an immediate win and an easy sale.

The other group is SMB/small enterprises that are experiencing scaling issues- specifically the exponential costs associated with vertically scaling a monolithic data store and/or API/serverless deployments. Our model is built to deliver a linear cost-to-capacity ration through horizontal scaling- if one node can handle 10k users for $100/month, then 2 nodes can handle 20k users for $200/month.

HarperDB was designed to be the first and last application platform you'll ever need, rather than swapping out your v1 database with something more capable as soon as you gain traction.

What are those kinds of guesses you mentioned when you realized you needed to change course.

Jack Boreham We knew we wanted to solve for throughput using a distributed architecture, but how users would interact with our product to provision it, configure it, and deploy it was based on our experience. When you then take that to market, you run into organizational policies and strategic relationships that require greater flexibility.

A great example is backup and restore. Our solution is snapshot-esque, but lots of enterprises are using infrastructure-level tooling that requires different operations in order to align with how those systems created and restored backups.

Another is our Custom Functions feature, which is our standalone Fastify server that lets you build your own lambdas. That wasn’t even in the original product, but we went to a large customer who needed the distributed database to live close to their massive serverless footprint… we thought, “Well, it can’t get any closer than building it in, so let’s do that.”

What're HarperDB's biggest milestones? Or the ones you're the proudest of?

Mónica Freitas, I would say our Custom Functions feature. It seems obvious now, but the complexities of distributed computing are more than enough to confuse customers not familiar with the concept, so anything we could do to simplify the infrastructure around it reaps massive rewards We coined the phrase “Collapsing the Stack”, and I think that is reflective of all the ways we focus on the developer experience.

It should just work.

So that’s what we focus on every day.

Those are very important aspects but aren't easy to put together. How do you build the perfect team? What factors do you pay close attention to when choosing the people that will help you grow and develop your idea? I assume this is a crucial aspect of the success of any startup

John T. You’re right. Ironically, this organization didn’t have a real clear policy on who to hire other than people we knew could do the job for which we were being hired. But along the way, we learned that for every 10x developer, you usually need someone with the personality to manage them to success.

We’re all working together, and we all have different skill sets. Everyone at HarperDB took the Meyers Briggs personality test, and the best-performing teams were ones where every person whose personality type ended in J was paired with someone whose personality type ended in P. This example worked for us and may not work for everyone, but it’s extremely helpful when you look at our little universe and how we find success with so few people trying to accomplish such a big thing.

That's great! I'm particularly interested in the small business part. What has the adoption of your product been for small companies? Have you noticed any specific challenge or need common among these businesses?

Cris Silva Small companies have been an interesting sales target. They usually come to us with a product that’s experiencing growing pains and are looking for a more scalable solution that won’t blow the budget down the line. The key, we think, has been getting developers to know and love us before someone in management tries to implement us from the top down.

Smaller companies put far more trust in their developers, and if those developers haven’t heard of us, it leads to more pushback than our self-service platform can usually overcome on its own… at which point we’re now selling it in with consulting and professional services, which isn’t our long-term strategy.

Can you give examples of companies that have used HarperDB?

How do you see HarperDB fitting in the metaverse?

Jack Boreham, I think the metaverse feels like an ideal use case for a distributed system… as it gains popularity, lots of metadata will need to be made available to a globally-distributed user base… and eventual consistency for things like presence or communication is a perfect fit for our platform. That said, there’s a lot to figure out… it’s certainly an interesting space to watch.

That's a fantastic feature. How long would you say it's the learning curve until a new customer can fully understand your platform?

Mónica Freitas Edge computing reduces latency by moving applications- and data, in the case of HarperDB- closer to the user. Moving APIs closer to the user with serverless tech like AWS Lambda lowers the time it takes to connect to the API, but if that API has to make a round trip to us-west-1, then latency is back on the table for someone in Sydney or Saigon.

Distributed systems lower cost by adding capacity along a linear cost curve (horizontal scale), as opposed to having to increase the resources for a centralized cluster (vertical scaling), which follows an exponential costs curve.

Lower complexity is achieved by abstracting away some of the connectivity logic between nodes, which is something we spend a lot of time on. We like to think we’re working on a platform that will change the way people think- or don’t have to think, really- about distributed systems.

Also, how can Edge and distributed computing improve latency and reduce costs and complexity?

Mónica Freitas The learning curve for HarperDB is extremely quick. Basic NodeJS knowledge is all that’s needed for Custom Functions development, but before that, imagine a document store as easy to use as MongoDB with a built-in http API to interact with it. Add our web-based studio that lets you see and query your data without installing anything at all, and it’s really the best developer experience out there.

Interesting approach 👏 at the very least, it's a fun team formation strategy. How big is your team currently? And how do you plan to expand HarperDB?

So you hope to have a fully independent platform without the consulting bits?

Cris Silva Ideally, yes. I like to follow the Zappos model. Their CEO famously said, “If someone calls the 800 number with an issue, we have failed, and should do everything we can to refine our process so nobody ever call that number.”

We work hard to make it so anyone can deploy HarperDB without reading the documentation.

How do you differentiate from current competitors in the data realm?

Mark Finnigan We think we have the best developer experience, we’re more often than not quite a bit faster, and we’ve made a few core design decisions to lower the total Cost of Ownership, like indexing everything by default, which reduces the need for a DBA to optimize your queries.

There are lots of players out there, though, so we accept that everyone has their best feature… we’re just betting that simplicity is going to matter a lot more than most of them seem to think it is.

Hey thanks for joining us here on Slogging!

I hope I'm not too late! My question is about startups vs corporations. You said, "I took a job at a fortune 50 company at one point because my wife was pregnant and I needed health insurance… longest 2 years of my life. Never again."

What did you dislike about the corporate world? What did you find most different about how careers work in larger companies vs the positions you've had in startups?

Hi Limarc Ambalina… I regret not being able to get to the last few questions, but suffice it to say, it’s exciting to have a chance to discuss the benefits of distributed computing architectures and our objective of making it easy to achieve for developers.

To answer your question, though, I like the immediacy of smaller companies. HarperDB Studio (https://studio.harperdb.io) was something I built version 1 of over a weekend. Our tagline is “Simplicity Without Sacrifice”, and as we strive to abstract away the complexity of distributed systems, my coworkers and I get to define the abstractions that accomplish that every day.

At my corporate gig, I waited 8 months for a single SQL Server to be provisioned to support an app that was saving the company $5MM/year running on AWS, but because we had a “competitive” cloud product, I had to provide it in-house… after 8 months I was pretty sure I needed to work somewhere “competitive” meant what I thought it meant.

Thank you for taking the time to chat with us. Any final thoughts you want to give our readers? How can people find you guys if they want to learn more?

Jack Boreham, thank you, as well!

My final thought is this: performance at scale is always a balance. That’s true for databases, and it’s true for companies. All you can do is find the systems and processes that work for you and your team and continually optimize them as you grow. It isn’t always easy or fun, but it doesn’t have to be as hard as some of us make it… humanity is a distributed system, after all, and the more transparent we are with information, the better off we tend to be.

Go spin up a free instance at https://harperdb.io, and build something that changes the world.