Cisco goes Christmas shopping, buys Cilium project originator Isovalent Cisco has bought itself a Christmas present: Isovalent, the startup that originated Cilium, an open source networking, observability, and security tool recently graduated to full project status by the Cloud Native Computing Foundation. As explained in Isovalent's "we're being acquired and that's awesome" statement, the project …
Alas poor Cilium, I knew him.
Cisco will try to cram a square peg into a round hole most of their backward customers will never ever figure out how to use while they still try to evolve beyond windows. As a Linux solution it's great, but as a Cisco solution it will be spurned and/or forked, thus marginalized as an undesired solution internally that it'll become another relic of the past like appdynamics, tetration, tail-f, viptela, and others.
Microsoft gave up a little bit before that, if you remember. Here’s a quote from an article on this very site, from 2019:
❝At Microsoft's Build conference last month, Gabe Monroy, lead program manager for the Azure Container Compute team, was asked whether Windows Containers are for legacy and Linux Containers for new projects. "I think that is a fair description," he said, demonstrating how the company's thinking on the subject has shifted.❞
Think of it: the coming of WSL2 is the beginning of a trend. Whether Microsoft is consciously planning it or not, the path of least resistance will be to rely on the Linux kernel more and more, while the Windows kernel gradually withers away.
We were talking about Docker. By 2019, the biz was not in a good place so I'm not surprised Microsoft was shifting its position.
C.
.... and for a good reason. Linux has a much smaller, and better designed, system API that it needs to support indefinitely. On the other hand, the necessity to drag Windows API *and* native (kernel) API to every new release is a severe drag on Windows development speed.
Dave Cutler’s original concept for Windows NT may have been a clean, modern design (for its time). But the original integrity of the design has become lost in a morass of so many short-sighted revenue-focused management decisions made since then. We see that in the deteriorating quality of updates and patches.
>> "and for a good reason. Linux has a much smaller, and better designed, system API that it needs to support indefinitely."
Another good reason is that Linux users, not marketing, drive the roadmap.
Container technologies (chroot, namespaces, control groups...), for instance, could never have been born at Redmond: they are all aiming at reducing the number of instances in datacentres. These capabilities are not exactly business friendly if your licence model is based on number of sold copies. Nothing Dave Cutler could have done about it (even remembering OpenVMS Galaxy).
Big things to note here is that Cilium and Tetragon are both CNCF products. So the Cisco acquisition nets them the Enterprise offering built on them and the top contributors, but not the core products.
It's a smart acquisition by Cisco. And likely the rapid pace of development on those products will slow depending on the truth of Cisco's new found "commitment" to open source. But the high quality tooling will still be there with support from all the hyperscalers.
Their current offerings of ACI and Talos will pick up some way better tech. The service mesh and cluster mesh already have applications outside of kubernetes which is core Cisco bread and butter. Combine it with their acquisition of Splunk and they have some great modern tech to integrate all together.
But like all of the big dogs acquiring to innovate and sell big enterprise packages. The just acquiring isn't enough. Take the Tetragon security acquisition, hope it's better than Palo Alto and their cobbled Prisma Cloud buying and poorly integrating TwistLock/BridgeCrew or Red Hat trying to slap StackRox into OpenShift.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
