Containers are the key to building that dream pipeline that will help you shake off your legacy applications and innovate to your heart’s content. Because that’s what cutting edge software development and deployment is all about, right?
But if you were attending DockerCon Europe in Barcelona in December you might have been taken aback by the amount of talk about migration and enterprise legacy apps, particularly of the Microsoft variety.
The vendor’s big announcements in the run up to the European conference were around its Modernize Traditional Applications programme for ageing Windows applications, and a tie-up with Mulesoft, again focused on legacy apps.
The cry from CEO Steve Singh was “no app left behind” with technical meat coming in the form of a spec for distributed apps with Microsoft and other enterprise stalwarts, and commitments on Docker Enterprise.
So does the bleeding edge look very much like the old edge? And is Windows the only legacy worth caring about? We sat down with Docker’s vp of product Banjot Chanana to discuss whether enterprises are hitting the pause button when it moves to updating and replacing their apps – or even reaching for rewind.
This future looks familiar…
When industry chatter is relentlessly focused on innovation, startups, and bleeding edge development, established companies often feel that modernizing their software development is an all or nothing choice, often leaving suffering paralysis when it comes to beginning much-needed migrations.
Docker’s role, according to Chanana is about “helping customers understand it’s possible and they have path for those legacy apps. Because most of them aren’t going to rewrite the monolith or try and modernise them on the fly.”
“In some cases they need to do simple things like upgrade the operating system and then they can think about next steps,” he says.
Once the legacy apps are containerised, the other technologies announced by Docker late last year come into play so they can be deployed on Docker Enterprise. Once this is all done, “You slowly figure out how to rewrite them, how to extract them into smaller microservices. That continues that path.”
Not that some elements in the business might not have already started without the organisation’s knowledge. Although this is often a known unknown.
“Over the last few months we have more customers because we’ve announced this program coming to us saying ‘I have legacy apps – I don’t even know if I have microservices yet, I don’t know if my developers are doing microservices, maybe they are, but the important things is I need to start somewhere and I need to do something with these Windows apps from 2008.’ It’s starting to shift the conversation.”
It certainly seems a more nuanced approach than simply telling established tech organisations: “You need to be like Netflix or you’re toast”.
And Chanana is upfront about the container pioneer having to do a little learning of its own when it comes to what enterprise customers want to do and are able to do. Four or five years ago, it would have been throwing all the hot buzzwords at its customers – microservices, CI/CD, DevOps. Rather a lot for a company to take in one bite, Chanana admits. “If you try and educate customers about how to do DevOps, how to be agile, how to build microservices, it’s a much longer involved process than just ‘here’s a piece of software’.”
Ain’t too proud to learn
The irony is that a major part of Docker’s early banner successes were at heart precisely the same thing: “I think when we saw those first successes with those first customers, we went back and said ‘let’s be a learning organisation, let’s find out how they did it.’ And they came back to us and said, ‘well we started with an existing app’.”
That said, Chanana is clear what companies should be ultimately aiming for – a common pipeline, through which different departments and different apps can move at an appropriate speed. “The stages can be different for every application, and every application can move through stages at a different rate. Legacy apps can be slower – monthly, quarterly, or only as needed – because there’s a lot of value there and you don’t want to upset it.
Meanwhile, “Your new microservices which you’re still building and learning, you’re updating those two, three times a day. And that’s ok.”
Not that multiple daily releases aren’t achievable – by some companies at least. “But there are very few companies at that extreme,” Chanana says. “And they’ll take a long time to get there.”
“It does take time to appreciate the reasons why organisations move at the speed they move”.
If it seems like much of Docker’s focus has been on Microsoft so far, Chenana says this is partly because the imminent pulling of support for Windows Server 2008 means there is a definitive deadline to aim for.
The companies have been working together since the early discussions about Docker on Windows in 2013, but in recent years it has much more contact with the Azure team, he says.
Moving off Windows 2003 and 2008 means not just moving OS, but onto Azure, and then exploring the wider range of services. For example, moving away from SQL running on a server under someone’s desk to Azure SQL. And of course, there is the potential for reducing datacentre costs. “It’s equally important to Microsoft to get their customers off their legacy operating systems.”
Windows and more
But Chanana insists Docker is not being drawn into a fixed orbit around Seattle, citing ongoing relationships with Amazon and Oracle amongst others. “Certainly our tools for MTA address not just Windows…but also support WebLogic, WebSphere, and Tomcat.”
“There are customers who in many cases are running large WebSphere clusters and it’s time to refresh those apps or change them, and the coupling between WS and the underlying OS and the underlying OS and hardware is generally very tight and very complicated and in many cases the administrators aren’t even around to help them.”
Converting to new applications has been a strong motivation on the Linux side too, he says. “Those applications have been hard to manage and keep up to date for many years.”
Over the course of 2019, Chanana said Docker’s major focus would be on the enterprise and strengthening the connection between developer teams, and the artifacts they create, and the operational and deployment side of the equation. “The other side…the artifact handover with CNAB, that’s going to be a strong focus for us.”
Needless to say, Docker will continue to add capabilities around Windows support with Chenani looking ahead to “additional automation capabilities, additional features on our platform to support Windows workloads, ensuring that you can run multi-tenant Windows workloads on the same platform.”
“There’s a lot of learnings over the last year and a half, we’ve incorporated all of our previous work with Microsoft to make Windows containers work on our platforms.”
“We’re continuing to add that support, so the upstream work of Windows with Kubernetes will be coming into Docker enterprise as well…And again we’re going to develop our learnings with Swarm, and networking capabilities supporting Windows, and we’re going to bring that over to Windows as well.”
So, if you’re part of of a forward thinking organisation, you may well see containers as being part of your future. But you might also be surprised just how familiar the rest of that future looks.