[Cryptography] Data in Use

[Phillip Hallam-Baker]

On Thu, Aug 4, 2022 at 4:14 PM iang <iang at iang.org> wrote:

> On 03/08/2022 19:31, Phillip Hallam-Baker wrote:
>
> So I got into a Twitter argument with Kenn White about the value of 'Data
> at Rest' security and I think I came up with a more general insight. In
> short, the traditional Data-in-Motion / Data-at-Rest diad is insufficiently
> descriptive. What we really have is a triad:
>
> Data at Rest
> Data in Use
> Data in Motion
>
> This gives us the acronym 'RUM'.
>
>
> Wow.  It's good.  How come nobody thought of that before?
>

As with 'Zero Trust', I am not the first person to use the term. There is a
Wiki Page.

Data in use - Wikipedia
<https://en.wikipedia.org/wiki/Data_in_use#:~:text=Data%20in%20use%20is%20an,CPU%20caches%2C%20or%20CPU%20registers.>

But this is talking about security concerns within a CPU which in the era
of SPECTRE, ROWHAMMER, etc, are certainly valid but I don't think it is
really getting to the distinction I am making.

To be clear, the reason I am making this distinction is because I have a
good solution for securing static blobs of data that does not apply to web
applications, databases and the like. And the more I thought about the
problem, the more I realized these are really two separate concerns.

The distinction also got me thinking about new classes of attack.

The Mesh can protect a confidential Word document on a USB drive. But that
is not going to provide protection if there is a malware virus in the
document that causes it to mail itself out to an attacker. Once Word read
the document, it was data in use.

There is quite a bit of writing on Data in Use on the Data at Rest Wiki
page as well. But it doesn't seem to have received the systematic attention
it deserves.

Another really useful aspect of this distinction is that Quantum
Cryptanalysis is absolutely not a concern for Data in Use. I am starting to
think it of less important for data in motion as well. If you think data is
so important you care about people possibly encrypting it in 2035 or
beyond, then you really need to think about data at rest and full end to
end security including storage. Sure, Mallet might collect data to decrypt
later but the sheer volume of Internet data makes that utterly impossible
to do except on a very limited scale.


If you go back Bruce Schneier wasn't the first person to say that security
is risk management, not risk elimination. But he was the person who
provided a book length systematic application of that principle. He
established the term in the field.

The only people who seem to have used the term are the folk looking at the
CPU level, hardware attacks and trustworthy compute modules. That is
important work that is clearly distinct from data at rest. But go up the
stack a bit to the database/Web application layer and we have a set of
attacks that are being considered as 'data at rest problems' because the
data is stored on the server but the reason the data is vulnerable is that
it is in active use.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20220805/102c81e9/attachment.htm>