TriggerMesh: Anywhere Serverless Integration

There’s plenty of hype around serverless technology being the latest hot thing, yet its growth ranks on par with the velocity of container evaluation and adoption in 2016, a recent Cloud Foundry survey reports.
A just-released report from Allied Market Research estimates the global market for serverless technology will reach $21.99 billion by 2025.
Cary, North Carolina-based startup TriggerMesh is focused on an integration layer for serverless functions and cloud native infrastructure enabling organizations to build enterprise-grade applications that span multiple cloud and data center environments.
In effect, it’s enabling all infrastructure to share their event data with serverless functions, and serverless functions and infrastructure to share their event data back to event buses.
“Think about it as something like Zapier or If This Then That. We listen to and filter event streams and then use that those builders to trigger functions,” said co-founder Mark Hinkle.
He likened it to Yahoo! Pipes, a service offered by the once-giant search engine company that allowed users to yoke together different tasks in a larger workflow.
“That is probably a good example of filtering all the data to get down to what you want so that you can trigger your function,” he said. “All of these event buses collect massive amounts of data, but having it codified in a single place that you can use as part of your cloud native event stack is not something that exists today.”
Integration Layer
Hinkle and TriggerMesh cofounder Sebastien Goasguen met at Citrix where they worked on Apache CloudStack. Goasguen went on to create the serverless framework Kubeless, a precursor to Knative, at a company he founded called Skippbox, which was acquired by Bitnami (now part of VMware), but Bitnami chose not to build out Kubeless to the extent he wanted.
Meanwhile, Hinkle became vice president of marketing for the Linux Foundation and later executive director of the Node.js foundation.
They founded TriggerMesh in 2018.
“We wanted to pursue serverless tooling and infrastructure that we thought would dovetail with where the industry was going,” Hinkle said.
Much of their early work was focused on Knative — TriggerMesh Knative Lambda Sources to provide cross-cloud triggers from AWS to functions on Knative-enabled clusters and TriggerMesh Knative Lambda Runtimes — but the company now is focused on providing an integration layer for serverless functions and cloud native infrastructure.
GitLab was the company’s first customer, for whom TriggerMesh built its custom serverless publishing platform.
GitLab was a one-off project, according to Hinkle, and the company is still building out its base technology with about 100 early adopters who are providing feedback on the product. It’s working toward having a preview product available by KubeCon Europe in late March.
Most of the early adopters are still working on a serverless strategy. They’re using TriggerMesh to publish an application and then they’re looking at how they may be using event sources to trigger things across their whole enterprise and to be more consistent with the way they do it, Hinkle said. They’re also looking at using serverless for stateful applications, rather than its initial stateless design.
The company recently raised a $3 million seed round led by Crane Venture Partners and Index Ventures.
“There are huge numbers of disconnected applications that are unable to fully benefit from cloud computing and increased network connectivity,” said Scott Sage, co-founder, and partner at Crane Venture Partners.
“Most companies have some combination of cloud and on-premises applications and with more applications around, often from different vendors, the need for integration has never been greater. We see TriggerMesh’s solution as the ideal fit for this need which made them a compelling investment.”
Redmonk analyst James Governor writes in a blog post that message buses will definitely be part of the serverless world, and there will be a world of triggers and events outside AWS Lambda.
“As we increasingly move into a hybrid, multi-cloud, multi-SaaS world, we’re going to need to bring together the new world of container-based functions and the kind of legacy event sources that IBM has long supported with MQSeries,” he said.
TriggerMesh maps well to the CloudEvents spec and also to the world Google plans to define with its AppSheet acquisition, he said.
“TriggerMesh is building message-oriented middleware for the serverless era, and enterprises using both on premises and cloud applications (which essentially means all enterprises) are likely to find the approach resonates,” he said.
“It will be interesting to see if TriggerMesh can build out a programming model that allows enterprises to take advantage of the service-oriented architecture work they have already done with platforms like MQ. If services are well described and managed, they can become accessible to cloud services. Some low code concepts map to this pretty well.”
Legacy Apps, Too
TriggerMesh provides a declarative API and tools to define event flows and functions that make up modern applications.
So far it has released EveryBridge, an event bus along the lines of Amazon’s EventBridge. EveryBridge consumes real-time data from event sources, such as VMware Vsphere, IBM MQ, Solace, or even custom applications and uses those events to trigger functions across various public clouds or serverless offerings hosted on Kubernetes.
“As an industry, people have jumped on serverless for greenfield applications and had really, really good success. Especially in that JavaScript Node community, those folks have written applications [and scaled them] for people like Uber and Twitter and Netflix,” Hinkle said.
“Now the traditional enterprise is saying, ‘You know, we aren’t greenfield. We have these applications that are taking up a massive amount of compute. And they aren’t Node applications, they’re Java applications and they’re Python applications and they’re other runtimes.’
“They have these legacy applications that don’t neatly fit into the Amazon, or Google or Microsoft Cloud as is, and so we’re providing integration tools that are going to allow them to take advantage of triggering these applications, should they say choose to move to a serverless paradigm.”
The company’s closest competitors are first-generation integrators like Mulesoft, “though what they’ve largely done in the past is integrate data and do more ETL,” Hinkle said, while TriggerMesh is going more in the space with companies like Zapier that’s creating interactions between the applications. Companies like Dell Boomi and Jitterbit fit in that space, though those companies were not built to be cloud native and are now working to become so, he said.
The company just released on GitHub three additional Knative event sources — IBM MQ, VMware vSphere and Azure Event hub Channel controller as well as a TriggerMesh Openshift operator.
Going forward, the company plans to add more event sources and is working on creating a simpler interface so users can add an event source quickly.
The Linux Foundation, Cloud Foundry and OpenShift are sponsors of The New Stack.
Image by ktphotography from Pixabay.