It’s a bad idea to bet against Finland, at least when it comes to open source. Despite Finland’s diminutive size, it’s given us some of the biggest names in open source like Linus Torvalds (creator of Git and Linux), Monty Widenius (author of MySQL), and Mårten Mickos (former CEO of MySQL). The Finns, in other words, don’t seem to know their place.
Aiven is no different. Started as an open source data cloud in 2016, Aiven offers managed services for a range of open source projects, including Apache Kafka, Elasticsearch, and Grafana. The company shouldn’t be doing well: Larger cloud competitors offer similar services, with sales teams that are more geographically proximate to customers than Finland-based Aiven. And yet, as Hannu Valtonen, the company’s chief product officer, told me in an interview, Aiven is doing really well, with data under management growing 5X in the last year.
How is Aiven managing this? In a very Finnish fashion: By doing cloud differently than larger competitors.
SEE: Cloud Data Storage Policy (TechRepublic Premium)
Standing out in the cloud crowd
Valtonen and I were talking about Aiven’s newest managed service, Aiven for M3—M3 being the time series database open sourced by Uber. Time series databases are hot because “the number of the metrics that you would like to collect has grown astronomically in the last couple of years, especially if you containerize applications,” Valtonen said. M3 had been on Valtonen’s radar ever since he heard about it at FOSDEM, but as Aiven customer needs for a robust time series database grew, Aiven decided to build a managed service around it.
This is fine, but hardly unique. After all, Aiven’s M3 move follows a number of other cloud providers–large and small–that offer managed time series databases. How does Aiven hope to stand out?
Unlike single clouds that operate multiple services, or single services that operate across multiple clouds, Aiven “operates multiple different services across multiple different clouds, and we do all of them with a twist.” The twist? It’s not actually one “twist,” but several, as Valtonen explained.
First, Aiven has been offering different managed services, like Apache Kafka, for longer than most others. Second, Aiven tends to stay current with open source releases, offering customers the most up-to-date versions, compared to other cloud vendors. But it’s the third “twist” that really sets Aiven apart, according to Valtonen.
“We have made completely different architectural decisions on how we create these services,” he said. Depending on the cloud vendor, you might be using Elastic Block Storage or Azure Blob Storage or Persistent Disk for persistence and backups. If a customer chooses a highly available service, the standard approach, according to Valtonen, is to give the customer the master server and then replicate the data on the block/disk level to the other availability zone. Aiven, by contrast, uses Postgres native replication, which allows them to use local SSDs for local disk storage, which, according to Valtonen, offers better performance. (Aiven provides its benchmarks on GitHub.)
Aiven’s approach also has it relying a lot on object stores: The company compresses and encrypts data before pushing it to these object stores. This allows Aiven to only pay for compressed object storage costs for backups, enabling them, in Valtonen’s words, “to have much bigger, longer backup retention for custom deals.”
I don’t know how this works in practice, but Aiven’s approach has enabled it to stand out in a crowded cloud market. It’s a very Finnish thing to do, taking a unique approach to open source contributions and competition.
Disclosure: I work for AWS, but the views expressed herein are mine and don’t necessarily reflect those of my employer.