1,027 karma · joined May 25, 2012
Software / Infrastructure Architect @ Udemy
Previously: Principal Engineer @ Microsoft Azure, Distinguished Member of Technical Staff @ Verizon, Master Engineer @ SUSE Senior Engineer @ Hewlett Packard Enterprise
ex - PTL for OpenStack Designate (DNS Project) (https://docs.openstack.org/designate/latest/)
ex - Member of the OpenStack Technical Committee
[ my public key: https://keybase.io/grahamhayes; my proof: https://keybase.io/grahamhayes/sigs/-W5bXnllXeYREBziC7KFYkLfh0caT8KLU6d7mQ20Tdw ]
I am still trying to figure out what superior features these are, even for me as a small sites.
For certs, I can have Lets Encrypt, and set a CAA record of letsencypt.org. For DNS, I can host my own, or use one of the myriad of providers out there.
I am not likely to have my domain taken away from me (and neither are most people), so the DNS ROOT switch is not a draw.
Why would I move to a DNS ROOT with no one on it?
We have seen this before, where people who run these services are personally held in contempt until they comply with the order.
Good luck with that. If there is no way to have a US court take ownership of a domain (or force someone to handover a bad faith registration), someone will either end up in jail for contempt, or it will just be banned.
> Additionally, HNS also fixes the problems the CA system has today. CAs can generate certificates for any domain - even if the domain owner didn’t request it. This has led to countless occurrences of abuse.
It fixes the non EV cert situation (kind of - CAA records also help this, and work on the standard internet), but doesn't fix EV certs (i.e. amzon.tld is actually Amazon LLC and not Scammers R Us Ltd.)
It also ignores the reality of most major companies in the world MITM'ing all outbound traffic from offices to ensure sensitive content is not exfilled from the company - there is 0 chance of them allowing traffic out that they could not MITM.
So every time I add a new LB, or cycle out an environment I have to pay someone? And then wait for them to process the chain to publish the result?
Remind me, what did etherium get to for transactons to clear? 20 something hours? That is not something I want when I am trying to redirect traffic from either a set of NS records that are broken, or if I host the full zone in the chain, a broken AWS / Azure region.
Combine the time to wait for caches to refresh on the internets (better these days, but still longer than people would like), with waiting for a blockchain to publish my new records, and you have ops people in pain.
DNS is already one of the largest fault tolerant, eventually consistent, cachable, federated, and globally distributed key value databases ever to exist - we don't need to add blockchain to the mix.
I get the "global, transparent, append only log" appeal, but for the reality of most people, I don't want to rely on a blockchain converging, or processing my update in a certain amount of time.
And that doesn't even cover companies (I don't agree with this, but it is a reality), that consider DNS entries private, and don't want to expose an easily scannable list of DNS entries .
If this was a commercial project, I would agree completely (and from a BOM cost, the manufacturer would be a lot better off investing the time in the dev on a microcontroller)
So in this case, if Bank of America had a company subscription with this hypothetical Maltese GitHub, BoA would be in breach, and could face sanctions (and this would probably cause the Maltese GH to get on the entity list as well)
Proving they haven't been caught, or may not even know about the US Dept of Commerce entity list.
So let us look at this:
1. GitHub becomes a Maltese company. 2. GitHub allows Iranian users to use all features 3. No US company could buy a GitHub Enterprise licence, or a private org.
The export restrictions don't let them notify people that the ban is coming - once they identify the users that are under the Export Restrictions, allowing them to download (aka export) their repos is a breach. They did as much as they could to allow them to keep their accounts at all (and this is only due to export restrictions having exclusions that cover "publicly available source").
GitHub did what they had to do to avoid being censured by the US Gov, and did it better than a lot of other tech companies recently, by leaving some level of access.
If you don't like the ban, and you are in the US, contact your elected representatives, and let them know - it is the only way it changes.
That is what they sell you alright, but actually getting access to these devs, unless you are a massive customer (and even then it can be difficult) you will never see anything more than a bug report, which might actually get more traction in an upstream bug tracker
A C(I|T)O wants someone that they feel that they can shout at if something goes wrong.
For you, with 20 nodes managed by yourself or a smaller team, CentOS / openSUSE / Debian / Fedora / Ubuntu all work.
(They also work for much larger situations, but RedHat / SUSE will sell execs on "fleet management tools" and "integrated installers" "SSO" and access to the devs that wrote the software running on top, and extract a fee. Most companies would be better employing a few highly skilled people, but that comes out of a different budget)
Even under Microfocus they enjoyed a level of freedom usually associated with independent companies.
If the board was left as is, and EQT is just a shareholder, that could be considered independent - more so than RH + IBM where there was some integration, and RH was becoming a wing of IBM.
But this is all speculation on my part, so YMMV
While there was a lot inside of SUSE that could have used more business focus, the ethos of the company is what made it so successful.
If someone is a Usain Bolt of software development though, they probably are someone who should be guided into, or are in, a senior engineering role. Glue Engineers / Smokejumpers / problem solvers are different, and useful, but when companies become over reliant on them it is an issue. And no matter how productive they are, they should be held to the same standards as other engineers.
The people/team handling it manually probably care
I would trust that a DE in GOOG knows how to leverage their team in the most efficient / useful way.
My point isn't that very smart / productive engineers don't exist (they definitely do), but that the meme of the dev in the corner who sits in the dark, and spews out golden code, without talking to anyone else on the team is actively harmful for our industry, especially in the smaller startup R&D teams that the current batch of publicity about them comes from.