There is actually a direct problem, the same problem most people have and do not know until it bite that they have. For almost all is named "cloud", for Science is named reliance third party data, publications, that today might be accessible from someone else computer, tomorrow we do not know BUT we might still need them tomorrow, especially data and publications about big and discussed things, scandals involving big enterprises, governments etc

Actually too many trust that what's available today will be available in ten years, but such trust in IT terms is named a big weakness, our entire history if only digital on someone else computer is at risk of being easy changed in the future. For science that need is as bigger and in theory scientist should be smart and acculturated enough to comprehend.

I can imagine, for instance that in ten years a scandal about a today new drug emerge, and in ten years original trial evidence is gone, on purpose and the scandal became then polemics just because there's nothing to really analyze so no one can really tell what happen.

I can imagine, in shorter time, that a day a scientist from a certain country found it's "own" paper collection along all it's notes, stored on some cloud reference manager gone, perhaps because his/her country have political disagreement with the cloud provider country or because it's university can't pay anymore access to such private and high-price platform.

If just we REALLY want peer review we can't even trust public places like HAL, we need to have data in our own hands, since they quickly became too much the sole alternative is having them on a distributed network where anyone hold some data and it's hard to believe a cartel to erase/temper them with a sufficiently large number of known nodes (like scientist authenticating themselves with GNUPG/PGP signatures with the relevant PKI infra and cross-signed keys behind.