Our vision for Tavros, “a cost-effective, cloud-native, integration platform composed of best-of-breed and seamlessly integrated open-source components”, started to form long before the 2020 global pandemic mandated such a platform. The first signs of Tavros started sometime around 2018 with OpenTracing (now OpenTelemetry).
With many decades of combined integration experience, we wanted to address the areas of an integrated platform where organizations chronically fell short at providing enterprise-ready software and alleviate the pain points. These included things such as observability of integrations for easier tracing and troubleshooting, integrated DevOps to aid in faster delivery, integrated configuration management to ensure deployability across all environments, Cloud Native Software to support ANY cloud provider, and a way to connect and transform data with ease and efficiency. These areas became the focus areas in which Tavros would be built.
Given the proliferation of the microservices architectural style, accelerated by the maturity and value of Kubernetes as a container orchestration platform, we were actively trying to find better ways to address the need for observability in our customers. We eventually found the Distributed Tracing (Dapper) whitepaper by Google, and the amazing open-source work done by Uber, Lightstep, and Red Hat, on the OpenTracing framework, all backed by the emerging Cloud Native Computing Foundation. We saw tremendous value and sound engineering that we wanted to bring to our customers. So we went to work and made our own open-source contributions in the form instrumentation for Mule and Apache Camel.
From this point we started tracking the expansion of valuable, independently developed open-sourced components addressing pain points we’ve identified time and time again in our customers, all growing organically with the support of the CNCF.
We started to have sporadic conversations like:
- X platform is really great at some functions, but not so much at others. It would be great if we could just plug-in component A for that.
- Could we somehow pick and choose the best piece for each concern?
- What’s the potential comparative return on investment vs established offerings?
- Can you imagine if we all had to stay at home indefinitely in a couple of years because of a crazy virus?
A couple of years later we were introduced to DataSonnet which was the “missing link” that was needed before we felt we could really commit to this effort. Data transformation is at the heart of enterprise integration (identify your bounded contexts!) We always keep an eye out for solid open-source alternatives to support our client base. With the introduction of DataSonnet, a data transformation DSL initiated by our partner ModusBox, in early 2020 we immediately saw the value in an expanded version of the tool’s capabilities. Based on the incredible work in Google’s JSonnet and DataBricks’ super fast Scala re-implementation we knew the engineering was solid and that ModusBox had already done a lot of the work in facilitating the extensibility needed. After a few months of engineering efforts, we co-released DataSonnet v2. In addition to the DataSonnet release, we constructed a native component to intertwine the language within the Apache Camel DSL. With a few tweaks, we were able to get inclusion as an official expression language within Apache Camel’s latest Long Term Support version 3.7.
With these two pieces figured out, observability and data transformation, the train started rolling. After 10+ years of enterprise integration work at hundreds of different customers and projects utilizing different technologies and components, we have a solid understanding of what supplemental tooling always proves valuable, which components are usually punted down the road that could otherwise accelerate time-to-value for a project; and the experience configuring those components! So we defined and committed a dedicated team to our vision.
Today, Tavros is an integration platform that not only compares to existing enterprise offerings but that we believe is better in many respects.
Here are some features to highlight:
- Provisioned and configured on-demand by Kubernetes Operations (kOps) and Ansible
- Infrastructure and Platform As Code (IaC), backed by GitOps and Continuous Delivery after provisioning
- Out Of the Box application Continuous Delivery pipelines
- Out Of the Box best-practices configuration for cross-cutting concerns like Logging, Tracing, and Mutual TLS
- Single Pane of Glass observability
- Many time-to-value accelerators for integration applications
- Single Sign On (SSO) experience across the supported components
- A Bring-Your-Own-Component design
- Out Of the Box API catalog and developer portals tied into the Continuous Delivery pipelines
- Commercial support available if needed
- Minimal to zero licensing cost
- No risk of vendor lock-in
The components that currently power the Tavros platform include: kOps, Ansible, Kubernetes, Flux v2, Helm, Kubeseal, Kong, Kuma, Keycloak, Jaeger, Elastic Cloud, Gitea, Jenkins, Nexus, Apache Camel, Spring Boot, DataSonnet, and OpenTracing. With plans for many others as part of our Bring-Your-Own-Component design goals.
With quick uptake already occurring by early adopters, we’re excited to bring this unbelievable amount of value to our customers, partners, and the open-source community in general. We hope to start engaging with the community to make Tavros even better.
For more information, check out our GitHub repository and our website.
About the Author: Jose Montoya
Tavros Principal Product Manager, Sr. Cloud Architect
As a Solutions Architect at MS3, Jose Montoya is in charge of delivering tailor-made software to our clients. Throughout his 5 year career with MS3 he has worked for multiple companies of every size and of various industries, always delivering business value by working alongside stakeholders at different organizational levels.
In his time off Jose likes working on community-driven open-sourced projects, staying up to date with the latest software engineering practices, watching sports, playing videogames, or tending to his garden in South Texas.
Education: Bachelor of Science in Computer Engineering from the University of Texas.