Serverless News - Talking About Planet Earth: Global Markets, Satellites and World News
Week of March 11
When researching for this newsletter, I saw a lot of news around various startups that specialize in serverless technologies. To me, this is a big deal. Seeing the phrase “Serverless” actually called out feels like a change of pace. Serverless is going from being seen as some behind the scenes glue to being seen as an actual enterprise solution.
#1: Ex-Uber Engineers create Fused, a serverless geospatial data startup
TechCrunch is covering this story about some ex-Uber employees starting a company called Fused.
Admittedly, this is not my space (pun intended) but I had fun reading about it. Here is the basic idea.
There are an astronomical (see, another pun) amount of satellites in orbit today. They are collecting an insane amount of geospatial data. This is data that references the earth’s surface. This data can be used for things like Google Maps but also in land surveying, monitoring forests, and so on.
It is really important for many industries to have reliable access to this data. Sina Kashuk and Isaac Brodsky helped build the mapping technology used at Uber so they are pretty much experts in this particular area.
They created Fused to help make the raw data more usable. It takes the from the satellites that are stored in a repository and then processes the data to make it more useful to the end-user.
Now this, in and of itself, isn’t new. There are many companies that do a version of this. What makes theirs different is the processing speed and simplicty. They are achieveing this by making their application serverless.
They’re goal is to do all of the processing in the cloud to speed up the processing capabilities. To do this, they will use a serverless architecture that spins up when processing then goes to zero when it isn’t. I think this will help them save money without sacrificing functionality or performance. They will also be able to pass benefits to their customers.
They have so far raised $1M in a seed round which is a great start for them.
I think this is a great story for serverless too. Don’t get me wrong, I think the real “win” for their product is the product itself. However, their ability to scale and serve customers with their geospatial data processing is because of their architecture (in my personal opinion). I have noticed an uptick in startups that brand themselves as a serverless solution for X.
For example, my friends at Nile recently raised an $11.6M round for their serverless Postgres startup. Neon Tech also raised a round last year. I expect to hear more about serverless data startups in the near future, especially with all of the talk related to GenAI recently. Pinecone, for example, now has a serveless vector database which are popular with GenAI developers.
I am excited to see the growth of Enterprise Serverless startups.
#2 Global Serverless Market was estimated to be $9.3B in 2023
Per DataIntelo, the Serverless Market in 2023 was about $9.3 and is projected to hit $41B in 2032. This ScienceDirect report suggests $36.8B by 2028 which is in line with the aforementioned projection.
This is a good thing as it lends credence to the idea that serverless can be used for enterprise applications. While there will certainly always be small projects here and there and startups working on their MVPs, a huge segment of that $36.8B.
One thing to point out in these reports is that the researchers conflate “serverless” with FaaS. In the DataIntelo post, they say
“Serverless computing, also known as function as a service (FaaS), allows developers to execute code…”
This irks me as serverless is WAY more than FaaS. With this in mind, the serverless market may actually be much larger. I couldn’t tell you how large it could be but I feel like you could confidently add 40% to that number at the very least.
There are many serverless container platforms for example. There is Google Cloud Run, Microsoft Azure Container Instances, AWS AppRunner, just to name a few. There are also a few Open Source projects that support serverless containers such as Knative and KEDA. You will hear me say this ad nauseum but serverless is WAY more than FaaS and we need to break this idea.
The reason this is significant is that it shows the evolution of cloud computing. In the early days of Cloud Computing, people simply saw this as a migration of VMs from On-Prem to Cloud. The idea here was to switch their tech debt from CapEx to OpEx. Instead of having to procure servers and house them, it just became a monthly bill of VMs that you spin up and down.
Time has progressed and now people are embracing microservices and now serverless to build applications.
#3 Guardian Build their “Pinboard” application using Serverless Architecture
I came across this story on InfoQ.
The Guardian created a new Newsroom collaboration tool using a serverless architecture. This tool is called “Pinboard” and uses AWS Lambda combined with GraphQL, written largely using React.
Tom Richards, software engineer at Guardian, had this to say about the project.
“For the past few years, a lot of digital communication in the newsroom had been happening over email and chat – crucially outside our editorial tools (tools being the place where the content people are communicating about is actually being produced) which creates the potential for confusion and mistakes given the pace of the news agenda. Meanwhile, there was a general feeling that pieces would benefit from preparing assets (images and video) earlier in the production process.”
I have never worked in a newsroom but I have a few friends in the press who gave me the short version of how things works. Newsrooms have editoral tools where the content is being created for hosting. A lot of communication has become digital and outside of these tools. The goal here was to centralize everything onto one tool.
One of the benefits of serverless is the fact that it is naturally event-driven. You will hear me talking about other benefits such as the savings with the pay-per-use model and the benefit of removing developer friction. But serverless apps are designed to be triggered by events. You can learn more about eventing at serverlesseventing.com (which is just an alias to my blog).
Reading this article, it seems like having event driven architecture allows them the flexibility to spin-up services as needed to allow teammates to collaborate not just in terms of communnication but in actual content creation. The assets get stored in a database and when another event occurs, the application spins up.
It looks like they are leveraging GraphQL as the API layer which is becoming a more popular standard for some API creations.
Pinboard is almost entirely serverless and is gaining traction within Guardian. There is already talk about adding more features.