@OtakarKlier wrote:
Hello @Brandon_Wertz ,
Would love to hear your thoughts on why and how to improve the design.
Regards,
@OtakarKlier -- I guess the title of the thread, to me, is mistitled. It should be "A Zero-Trust Strategy incorporating Prisma xxxxxxx" For me, given that we're starting off on the wrong foot you're not going to achieve the outcome you're looking for.
My interpretation of ZTNA's intention is to really segment and isolate the enterprise as a whole, while creating specific granularized (authorized) access consistently throughout your enterprise no matter where you are both logically and physically. You want to create a consistent user experience that no matter where the user is, the user accesses resources the same way and the support and oversight of the technical components are consistent for the IT and SecOps support teams.
ZTNA, is the realization that cyber exploitation / compromise in a legacy network design is a foregone conclusion and networks/applications need to mirror to a degree the controls that exist in more secured networks like PCI or the military.
The above stated, if someone is going to leverage Prisma Access Mobile Users (VPN) that's one external component, are there others? Is there a VDI solution? What's that OEM? What controls exist there? What does access from the internal network look like? What visibility is there? What controls exist? Are they the same as when a mobile user? If as MU you're super locked down with great visibility, but internally you're wide open, with little control and near zero visibility with visibility just on what's happening externally then that's not really a ZTNA paradigm. At all layers of your network you should have consistent application and use and user based controls, business application based access restrictions as well as some level of network access control (NAC) employment.
To do this you're going to want components that integrate with each other sharing data, visibility and functionality (if possible.) A single pane of glass into your entire enterprise. So if you block or allow something that action should be executed in one location (view) with the action happening in maybe 4+ separate network segments/layers.
... View more