• Album
    link
    fedilink
    10
    edit-2
    4 months ago

    People are definitely talking about that. Maybe not in the media but in the affected industries.

    It was different when we weren’t living in a software as a service world. You could be mono platform but since you had complete control you didn’t have to worry you could roll out how you wanted.

    Crowd strike by it’s very nature is supposed to be live updated throughout the day as threats emerge.

    If we want services like these maybe we need to come up with better ways to isolate them from the kernel while still allowing crowdstrike type software to detect threats

    Solving the crowdstrike problem could solve kernel level anti cheat software too

    • @ramble81@lemm.ee
      link
      fedilink
      2
      edit-2
      4 months ago

      You’re talking about the IDS/IPS problem. For it not to impact the kernel it would need to be a passive, read only system. But if you need it to be active to actively prevent threats it needs to have the same level of access a threat actor could gain. You can’t move everything to user space without a shit load of signing and things like TPM and SecureBoot which people have been decrying for years as “vendor lock in”. So at some point a level of trust or risk must be accepted.