问题原文
原帖链接 · Sergey · CC BY-SA 3.0
What's the best strategy to keeping access tokens secure in the event of a compromised server?
Let's say I have a Linux web server to which I regularly ssh into. There is a variety of critical access tokens stored on this server (e.g. encryption keys, Amazon S3 for backups, etc). If an attacker can ssh into the server, what's the best strategy to prevent them from reading the keys?
已采纳解答
原帖链接 · Jeff Ferland · CC BY-SA 3.0
It's pretty hard to keep keys from being read. Mandatory access control on your daemons using AppArmor or SELinux will kick up the ability to limit read access.
Making use of software like Vault kicks it up a notch. Like the temporary credential system given by Amazon's STS, this can be applied to any backend you write a module for a few exist thus creating temporary and easily revocable credentials. If written directly into your app, this will keep passwords outside of configuration files.
You can use an authentication backend such as the App ID one or if you have hardware, get super fancy and write a TPM based one.
摘自公开归档,保留原文语言,版式经整理,外部媒体与邮箱地址已移除。
