Hacker News Reader

Comment by heybrendan | Hacker News Reader

Posts like these always seem buried in deep marketing lingo. Can anyone "in the know" ELI5 how this is different than Google Anthos? [1]

If they are in fact similar, I can't help but be left wondering why ACG didn't devote at least a sentence or two to comparisons that already exist in the industry.

[1] https://cloud.google.com/anthos

View on HN

The post does mention Azure's Arc and Sentinel, but definitely should have called out Anthos specifically.

The interesting part of this to me isn't so much that other solutions already exist -- Azure and GCP, due to their catch-up position, have always been far more conciliatory toward the reality of other cloud workloads than AWS. It's that AWS is signaling, however grudgingly, that they will have to play ball with other clouds in order to stay ahead. That's a big shift.

Thanks for pointing that out. Unfortunately, still no mention [in the post] of how Azure's Arc/Sentinel is different than what AWS has seemingly built here. And agreed, it's an interesting signal.

Hi Forrest, hopefully constructive feedback - your choice of font color (rgb(74, 93, 133) I think) for text overtop #fff background is extremely washed out for me, not quite unreadable but heading in that direction. I might suggest a darker choice of font color for better readability, I just gave up trying to finish reading the article due to eye strain.

Edit, screenshot: https://postimg.cc/kRmGmSJh

Thanks for this feedback - will share with the ACG design team.

The difference is (just based on what AWS say about ACG) is that anthos comes into its own when managing multiple clusters. The drawback is that all the most common infrastructure requires you to be all-in on GKE and the google stack. If you want to use Cassandra or Postgres, for example, you either manage it yourself or use the google options. The problem with this is that you basically give up all the benefit of having a multi-cloud approach by following this approach. However! If you have a bunch of legacy systems sitting on an on-prem set of self-hosted hardware finally you can mix and match between the on-prem stuff and the GCP managed cluster.

ACG sounds like it can be used to manage hosts and containers; so it's more like enhanced k8s / EKS than a meta layer above kubernetes. This is just my take on this product based on how the marketing lingo talks about it. I'm much more familiar with AnthOS

Reply on news.ycombinator.com