4 ms·
The Case for Centralized IAM (2022)
- bob1029 3y ago> Having the code of the authorization layer mixed in with the application code itself can be very problematic. Most importantly, it creates a situation where we struggle to upgrade, add capabilities and monitor the code overall as it is replicated between different microservices. My takeaway from this article is to avoid using microservices.
- chayesfss 3y ago[dead]
- crabbone 3y agoI had a coworker who used to say Tight coupling is good because it solves communication problems. No coupling is good because it solves communication problems (but worse). Loose coupling is the worst deal you can have. This is in the context of microservices which advertise themselves as loosely coupled (as a system / automation person I can highly relate to this sentiment). I think that the best project management strategy is to hold out for as long as possible before splitting a program into multiple parts, and once that becomes inevitable -- split off completely into separate programs. The intermediate state of having many components that depend on each other a great deal is the worst possible situation you can find yourself in as a developer. Nothing ever works properly, the staging and the testing systems are always in some kind of "dirty" state that doesn't reflect the intended operational state. There's a ton of automation that tries to keep components up to date with modifications in other components and invariably fails due to its own complexity + human factor.
- butterisgood 3y agoMonoliths have advantages except in that they also invite complexity (they do not necessitate it) when there's no clear interface between components. Even with interfaces, any empirically observable behavior can become a kind of contract - so nothing is perfect here. Microservices seem to serve Conway's Law more than anything else I can understand. Groups want to have scoped projects, and a bubble within which to work and validate what they do. But microservices seem to potentially hinder system-wide traceability, system-wide debugability and performance. Modularity eschews one class of problems for another. Both monolithic and modular designs require discipline to keep things manageable, and they're not the same kind of discipline or maturity.
- crabbone 3y agoThis is not an argument against modularity. The argument is that if you want modularity -- go all the way. Microservices is the argument that says "nah, let's divorce, but keep living in the same house". Here, to make it more concrete: two Unix programs assume no knowledge of each other's existence (ideally), they rely on standard input / output / pipes / signals (i.e. some communication methods that aren't unique to any of the programs, but provided by the third program -- environment). This, in my example would be to go all the way: no need to know that the other program exists. Microservices are as if you had an OS similar to Unix, but any pair of programs could decide on their own kind of pipes / signals / the way loader resolves symbols / function calling convention etc. Now, the argument the proponents make is that it's hard to make a universal "medium" used by all programs run in such an environment, and that if you want to "move fast", then you can skip working on this universal "medium" and just hack it together for any pair of components you have in the fastest way you can. Despite how ridiculous it sounds... this argument isn't without merit. I've seen a lot of systems that tried to develop "universal medium" and failed. The latest largest such example is Kubernetes itself, if you think about various entities in Kubernetes as being components that need to communicate to each other. Kubernetes developers are slow to acknowledge the failure and are trying to fight it point-wise (eg. not so long ago they axed PSP), but the system keeps getting more complex and unpredictable with every new edition. So, in frustration, one may decry the idea of universal medium, suggesting reducing the scope of the problem... and that's what microservice apologists seem to be suggesting. But, to me, it sounds like giving up on an essential desirable quality of a program. Kind of like if you wanted an automobile, but you couldn't get the engine to work quite right, and, in the end, abandoned the effort and just rode a horse.
- chayesfss 3y ago[dead]
- mberning 3y agoIt’s an interesting idea, but I don’t see how this is really going to simplify the life of the IAM team in an enterprise. Even if you use their proxy to presumably add authorization capabilities to a legacy app, there is still some type of auth model in the app itself which needs to be contended with. At the end of the day I wish SCIM were better and more widely adopted as it makes the need for tools like this go away and lets any idp, iam, or iga tool integrate with any application.
- Loic 3y agoSCIM: System for Cross-domain Identity Management https://en.m.wikipedia.org/wiki/System_for_Cross-domain_Identity_Management https://en.m.wikipedia.org/wiki/System_for_Cross-domain_Iden...
- 1letterunixname 3y agoYep. 0. Get your SSO + Webauthn 2FA everywhere. (Don't use Okta or Auth0! Do it yourself!) 1. Have service accounts that are owned by teams. (Every team should have an "email" or group chat address, and use that as an identifier everywhere for who owns what.) 2. Have a proxy middleware that requires current good telemetry from regular client security checks. (Deploy some sort of firehouse structured data streaming service, say fluentbit on the managed hosts.) 3. Centralized intranet web platform that's effectively Istio pointing to micro-fullstack-ends. With a globally-routable, VPN-less domain. ("VPNs: Booo!" But if you need one, use managed Wireguard, not some ancient Cisco box no one wants to touch.) 4. Have an internal wiki. 5. Have a (de-)centralized search engine. 6. Roll machine & user (intranet & WiFi) certificates every week or 2. (Lock machines if they haven't checked in in 4 weeks.) 7. In offices, deploy 802.11x and POE+ every-frickn-where. 8. Don't do BYOD or personal apps on corp. Legal and support disaster areas. 9. Deploy Yubikey. 10. Get a password manager. Keeper scales well. 11. Throw away Zoom, iMessage, Telegram, and WhatsApp immediately. Use Session or Signal. (Solving mandatory min/max retention limits is an exercise for the reader.) 12. Disallow many cloud features clearly. 13. Have a 3rd-party approval process to run it by your legal, privacy, and client teams. 14. Retain the ability to remote insta-lockout and remote wipe machines. (And record any exceptions to legal holds.)
- phrotoma 3y agoI have an intuitive sense of each of these except 2, could you elaborate?
- mschuster91 3y ago> 0. Get your SSO + Webauthn 2FA everywhere. (Don't use Okta or Auth0! Do it yourself!) Maintaining that crap is a full-time job in itself and very hard to get right.
- tecleandor 3y agoI guess that, unless you have a small number of users, you could buy a whole authentication company for the price of your OKTA licenses...