UPDATE – 20 Nov 2018 – I wrote this back in 2016 as I was just experimenting with Docker. Since that time, not only did Swarm emerge as Docker’s tool for container management/orchestration/clustering, but we also saw the emergence and then domination of Kubernetes as a tool for container orchestration. I’m leaving this post online, but at this point the examples are quite prominent for how Swarm and other tools can be used.
Do you use Docker Swarm? If so, how? I have been incredibly intrigued ever since reading about the release of Docker 1.12 earlier this week.
As Benjamin Wooten writes, now with only two commands:
- We get a deployment platform which gives us resilience, robustness, failover and fault tolerance for our containers.
- We get load balancing and a routing mesh which makes service discovery simple.
- We can use our server resources more efficiently with various allocation strategies.
- We can scale containers up and down with one command.
- Communications within the cluster are secured with dynamically rotating certificates.
Ever since, I have been reading more, such as this piece about setting up a swarm with Raspberry Pi systems.
Now I am curious… how are any of you reading this using Docker Swarm? What are doing with it? I am intrigued and curious to do more…
What does Amazon.com do so much better than Google? And why does Amazon do everything “wrong” while Google does everything “right”… yet offer a better platform? How should you construct a “platform” so that everyone can use it?
If you are a developer, IT manager, product manager, system architect, product marketer, CTO or even a CEO, you really need to take a bit to read this “Mother of all Reply-All failures” that was written by Googler Steve Yegge and accidentally posted publicly back on October 12th. Steve pulled down his own posting of the rant, but it was re-posted to Google+ by Rip Rowan and also posted over to Hacker News. The long rant – and the comments on both sites – are worth a read:
It’s a LONG piece that gives some fascinating insight into both Amazon and Google as companies, but also into what it takes to be a “platform”.
A bit later, on October 21st, Steve Yegge posted an update indicating that he did not get fired and in fact people actually listened within Google. He also dove a bit more into Amazon.com and Jeff Bezos. And just this week he wrote a lengthy piece describing how amazing it is to work at Google, explaining a bit more about what he meant in his rant about how Google “does everything right”.
The original platform rant, though, should definitely be on a “must-read” list for people thinking about how their services could really be a “platform”…
P.S. Are we connected on Google+? If not, you can find my Google+ profile and add me to a circle…