Cloud Native Computing Foundation KnativeCon EU 2022, 19 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Lightning Talk: Modernizing Your IBM-MQ Applications with Knative and Kong - Sebastien Goasguen

Description

Lightning Talk: Modernizing Your IBM-MQ Applications with Knative and Kong - Sebastien Goasguen, TriggerMesh

IBM MQ has long been at the core of many enterprise applications especially those using mainframes. Enterprises are trying to extend the lifetime and relevance of such legacy systems in a Cloud-Native era by linking them with containerized workloads. Legacy applications still heavily rely on these systems and second, they cannot be easily replaced by Cloud solutions. In this talk we will show how you can bridge the two worlds of old legacy systems and modern Cloud architecture. We will make use of a Kong API gateway, Knative and some TriggerMesh components for event transformation and connection with IBM-MQ. We will take the example of a Mulesoft application that exposes a REST API in front of a mainframe system of record and we will decompose it using an event-driven system built on containers and using Knative and TriggerMesh. This will show how you can keep using your IBM MQ system but usher into a more open source and cloud-native world which speeds up your application workflow and reduces cost.

A

Thanks carlos uh hello, again everybody so my name is sebastian goes again. I'm the co-founder of trigger mesh been involved with uh k native for quite a while, and I got five minutes to talk to you about this. Modernizing ibm, mq application with k, native and kong, don't add kubernetes because it makes for a very bad acronym.

A

So we'll we'll stick with k native and kong. It's a very interesting use case. At the same time, it's a bit of a frankenstein.

A

So actually, once I'm done talking, I would love to hear from you what what you think, okay, so what I mean by modernizing applications. You know what what we're seeing when you know talking to uh you know: customers, prospects and so on. It's you know they want to move to uh to open source. It seems uh you know it seems rather simple, but it's actually, you know a big, a big desire right.

A

It's about removing proprietary software as much as possible, and the reason is pretty simple- is to cut costs all right, so adopting open source as much as they can and while, while they do this, they need to integrate with existing systems that they cannot get rid of right. So big proprietary systems, like you, know, oracle db, uh ibm mq, you know you name it right. So it's this!

A

You know this desire that they have to uh to become less dependent on very expensive systems, adopt open source, uh but at the same time they still need to keep on on using. You know the infrastructure that they have right, so move to open source still integrate with the existing system, make them relevant.

A

As you know, they move to the cloud uh as they adapt cloud cloud native technologies, cloud native techniques you know and uh and so on, right um and then the the the third you know, big thing about modernizing application is really to you know: rejuvenate the applica, the overall application and operational landscape, and it's more of a you know, culture thing right and we've been talking about in devops for a long time.

A

It's you know changing the culture of the enterprise. How do you deploy apps? How do you maintain them? How do you monitor them and and so on? So it's pretty, you know it. It seems simple, maybe here where we are very much cutting edge, but there are lots of enterprises out there that you know for whom adopting githubs, you know is quite challenging and it is a big change because that's not how the the deploy application uh monitoring with new things like prometheus.

A

It's a big change right, uh but you know nonetheless the they all wish that they would go to prod much faster right. So you know that's that's the thing right. It's changing the culture changing a little bit of the landscape so that you know they go much faster and and they reduce cost. You know it's as simple as this: that's the that's the benefit of of modernizing.

A

I don't know why it's flickering, but okay, so big boring use case right. So lots of lots of apps that we're seeing in you know, enterprise that we're talking to it's uh lots of java applications, and you know a lot of them. Look like this, so it's http, you know uh http server, returning java, exposing a rest api and behind it you know you're, actually uh fronting you know, mainframe applications with whom you're communicating with using ibm mq right cannot be more.

A

You know legacy or enterprise than this right and you can have thousands of those apps right. So you know if you use something like mulesoft, for example, you've written you know a rest api with mulesoft and then you're doing some data transformation in java and then to communicate with your mainframe you're, going in and out of mq ibm mq right uh and in and out of mq it's. You know, as I said, to talk to you, know very old system of records running on on mainframe.

A

So that's where we start from right, and how do you modernize this? You know which path do you take and so on? So what we're seeing is that you know it's a little bit why I mentioned you know the api gateway earlier in the in the panel.

A

So we are seeing you know, adoption to newer api gateways like kong right, so we're seeing lots of enterprise moving to kong and then, if you start putting kong in the picture- and you imagine that they're trying to rejuvenate or modernize the you know the platform on which workloads are running, uh which means kubernetes.

A

Suddenly you end up with you know, kong as a gateway uh wanting to deploy in kubernetes, but still you have to deal with your old system of record and your mq right. So how do you you know? How do you get out of this? How do you strangle this and uh and and get moving right?

A

So that's where I'm actually very happy about this, but it's also a little bit of a frankenstein from a you know, technical standpoint, but uh you know I so I would love to to to see what you what you're thinking about so we cannot get rid of. The system of record is still in place. You're still going to talk to it through mq right and then suddenly you have kong, that's exposing your rest, api right!

A

So in the middle you know we're now putting trigger mesh here and with our apis. What we provide is data transformation, so all the transformation that you could do in plain java, you can do it declaratively with canadian objects. For example, you can do filtering, you know, processing with functions. You know whether you use our functions or kn doesn't matter and then to talk to mq.

A

You use well an mq source to read from a queue, and then we have an mq target so that you can put into a queue right so a source and a sync into into mq right. So if you start, you know, you start from this and suddenly you move to something like this that moves that deploys in kubernetes is defined by api objects. So it's fully declarative right and you can drive it with ci cd right. So it's great! It's very modern!

A

That's where you want to go, but you know what we're finding is that the the path to go from the old to the new? That would look like you know. Something like this, you know, can be a bit challenging because it's a it's a big step.

A

So the step is that you know suddenly your kong as an api gateway. That's your entry point and what's behind you know, I showed you an example, but it could be much more asynchronous and it could be an event-based flow. So your cong rest api suddenly now needs to bridge to an event-driven flow right, so you're you're, bridging two worlds here and kong needs to emit cloud events right. If you want to you, know uh deal with uh with k, native and uh and eventing constructs and and so on.

A

The flows can be asynchronous, but if you want to reply to the client with a you know a reply at some point, you need to synchronize all right, so we've built a synchronizer using you know a correlation id in the events so that we can wait for a particular event back right and then the final step is that we need source and target to mq. That's actually the the easy part, but we've done it.

A

You know as canadian objects, so everything here is available on github trigger mesh kong, ce demo c for cloud event right so and you'll get the the full flow. So the first thing is a a cong plugin that uh transform a rest request into a cloud event right, so you curl to kong. You give it some some payload and suddenly you have a cloud event. You know thanks. So kai thanks scott, so you gotta, you got your cloud events here.

A

uh You know emitted by uh by kong, which means that suddenly you go rest and now your event into a communities cluster, all right, so you've you've made that path.

A

This has the advantage that, depending on the route that you hit on your api gateway, you could have different cloud event types and you could now do routing based on event types right. So we fall back on. uh You know an eventing type type world, uh yeah yamo, so mq connectors.

A

You know, as I said, uh that's the easy part right. So we have brand new api kinds, uh mq source mqtarget, you specify your credentials, you specify your queue manager, your q name uh here on the target. You know: that's that's something. That's interesting with mq is that if you put a message into a specific queue, you can say hey, the reply is going to be in another queue right. So here you know it's the it's the request, reply mechanism.

A

So here you have a reply to parameter that says: hey the mainframe system of record is going to put its reply into this other queue, so read from that other queue to get the reply all right.

A

So that's your your connectors to mq, now open source, fully declarative, you're, good and suddenly the flow looks like this.

A

uh You end up with talking to kong, open source or getting uh you know the enterprise uh product from them right, but you could go fully open source with the kong api gateway. You use our cloud event. Plugin you get an event into a native router. You can do routing based on the event types.

A

uh Different transformation function based you name it. How you want to do the transformation you put into mq your mainframe could do its job put the result back into the output queue which we read with our source and then the magic which is a little bit. Frankensteinish is the synchronizer right which is going to wait for an event with a particular id to make a synchronous reply all right, so we have an asynchronous flow. uh You know that we are exposing with a you know, a synchronous mechanism.

A

So that's one way of modernizing you know such uh such a flow here uh and just to you know, just to go a little bit. You know yeah, I'm the guy who is not using dark mode.

A

So I'm just going to.

A

So I don't know if you can see this, but so a trigger mesh with you know to make things a little bit easier. We've written this integration, language which is based on hcl.

A

So here you see that entire flow that I described with our language with our router so based on the event type, so route attributes you see the different type go to you know a specific transformation transformation could be function here. That's our! You know function example and then, at the end I mean you see that a function can say hey once you've executed. This function go to mq, so it says you know to target mq and you have the definition of the mq target right. Of course, this is not.

A

uh You know, kubernetes object, but we can generate you know from this. We can generate the yaml. That's where the developer experience, I think you know to me- gets complicated, because if you ask a new user to start writing a flow like this with pure yaml, you know it becomes very complicated, but here you see the full yaml that gets generated with you know the source to mq uh the synchronizer.

A

uh You know the different triggers the different functions right, the broker, so you get all these these objects and the compilation of all those declarative objects makes your and your entire. You know flow.

A

All right, so, thank you so much modernizing your apk, your mq applications with kong and k native. Thank you.