Re: [Fed-Talk] Š. (was: CoreCrypto / CoreCrypto Kernel now in "In - Review"(CMVP))
Re: [Fed-Talk] Š. (was: CoreCrypto / CoreCrypto Kernel now in "In - Review"(CMVP))
- Subject: Re: [Fed-Talk] Š. (was: CoreCrypto / CoreCrypto Kernel now in "In - Review"(CMVP))
- From: "Marcus, Allan B" <email@hidden>
- Date: Thu, 21 Feb 2013 17:07:53 +0000
- Thread-topic: [Fed-Talk] Š. (was: CoreCrypto / CoreCrypto Kernel now in "In - Review"(CMVP))
No, only the data in the Good sandbox or Apps in the Good Dynamics
environment is protected. Any data in any other app that doesn't use Apple
Data Protection (ADP) to encrypt the files is not protected and can be
accessed before the user unlocks the device. Also, Good's UI is not nearly
as good as Apple's UI, which goes to my point that Apple is destroying the
user experience by forcing those that need a higher level of data
protection to use products like Good. Also, the data protection for data
in the Good sandbox and data in other Apps that use ADP is protected with
different crypto modules.
--
Thanks,
Allan Marcus
Chief IT Architect
Los Alamos National Laboratory
505-667-5666
email@hidden
On 2/21/13 8:49 AM, "Villano, Paul Mr CIV USA TRADOC"
<email@hidden> wrote:
>Allan I don't understand all of the angst you're perceiving or why you
>think
>your solution is something new. What you describe sounds like what the
>Army
>is doing with the Good(e?) solution. The whole reason I understood they
>were
>chosen was because they follow the paradigm you mention of protecting all
>of
>the data on the phone, not app by app. What am I missing?
>
>-----Original Message-----
>From: fed-talk-bounces+paul.villano=email@hidden
>[mailto:fed-talk-bounces+paul.villano=email@hidden] On
>Behalf
>Of Marcus, Allan B
>Sent: Thursday, February 21, 2013 10:31 AM
>To: email@hidden
>Subject: Re: [Fed-Talk] Š. (was: CoreCrypto / CoreCrypto Kernel now in
>"In -
>Review"(CMVP))
>
>Again, thanks for the taking the time to answer. I'm not going to let this
>drop. I also have to admit that I'm getting frustrated with the pedantic
>nature of your replies.
>
>As for being pedantic, take for example, your comment regarding login vs
>unlock. Wikipedia describes login as
>
>
>"In computer security, a login or logon refers to the credentials required
>to obtain access to a computer system or other restricted area"
>
>
>I submit that when a user takes an iPhone and enters a password to unlock
>the iPhone, she has done the above. As the iOS device is not a multi-user
>device, there is not a username to identify the user, just the password as
>the credentials to obtain access to a computer system. Based on the
>definition you gave, the user is logging-in as well as unlocking.
>
>I'm going to rant a little now, but I think what I'm saying is fairly well
>agreed upon by anyone on the federal government providing iOS devices.
>
>When talking about working with developers "given the need to mitigate
>these risks and what is available today", I submit the most practical
>solution for the entire federal government is to pressure Apple to provide
>for a complete data protection solution. Any other approach is essentially
>futile. You also spoke of the capability to leverage third parties to
>provide solutions, but IOS is different from Mac and Windows. Desktop
>operating systems are not nearly as locked down as iOS. That control,
>which
>Apple insists upon, also puts more responsibility on Apple to provide
>solutions, since third parties cannot. Whole device data protection (for
>lack of a better term, and not wanting to trigger your Pavlovian response
>that the device is already encrypted) is not a solution a third party can
>provide on iOS (as they can on OS X). The ONLY provider than can provide
>this is Apple.
>
>
>You seemingly managed to ignore my comment
>
>
>
> "When I say I want a global/system solution to turn it on, I mean in
>the System App. I wan enhanced data protection for all apps, all the time.
>Yes, I want the user to have to enter a password to do anything on the
>device except make an emergency call (or a regular call by entering the
>number directly). "
>
>
>
>I will refer to this option as "whole device data protection (T)." WDDP
>is a
>trademark owned by Allan Marcus and can be licensed for a small fee :-) If
>Apple implemented it, I will sign over the trademark to Apple for no fee!
>
>
>The only vendor that can provide this is Apple. I'm sure you have heard
>this
>request before. You state Apple "we will not destroy the user experience".
>For those of use that comply with the FISMA law and use Good, don't you
>think that Apple's lack of whole device data protection has force use to
>"destroy the user experience" by using Good?
>
>Also, please take a few minutes and explain how using data protection for
>all data on the device will "destroy the user experience". I assume some
>things will not work, like certain background syncing or processing, but
>if
>there were a Yes/No slider in the general System panel to turn on "whole
>device data protection", the user could easily be warned that certain
>things
>would not work with that option. Apple allows the user to turn off
>cellular,
>bluetooth, wifi, and a ton of other settings that could "destroy the user
>experience".
>
>
> For example, I submit my iPhone SUCKS when I turn on Airplane mode!
>Why the heck does Apple DESTROY my user experience when I turn turn on
>Airplane mode! My device's user experience is also destroyed when I turn
>off
>siri. Why does apple allow that, but not whole device data protection?
>Apple
>let's me exclude certain data from spotlight searches. I submit that when
>I
>exclude mail from spotlight searches and then I search and forget that
>I've
>turned off mail and no mail comes up in a search, my iPhone is _broken_.
>
>
>Obviously these are ridiculous assumptions to be made on the part of the
>user. These are options that the user can turn on/off, and once on/off the
>user should not expect those options to function. The same can easily be
>said for whole device data protection option. Let the user decide to give
>up
>a little functionality to gain the security.
>
>I realize you cannot talk about future features. What I would like to hear
>is who do I have to talk to for this feature to be seriously considered
>for
>a future iOS?
>
>
>--
>Thanks,
>
>Allan Marcus
>Chief IT Architect
>Los Alamos National Laboratory
>505-667-5666
>email@hidden
>
>From: Shawn Geddis <email@hidden>
>Date: Wednesday, February 20, 2013 4:21 PM
>To: Allan Marcus <email@hidden>
>Cc: "email@hidden" <email@hidden>
>Subject: Re: [Fed-Talk] .. (was: CoreCrypto / CoreCrypto Kernel now in
>"In -
>Review"(CMVP))
>
>
>Allan,
>
>On Feb 20, 2013, at 4:48 PM, "Marcus, Allan B" <email@hidden> wrote:
>
> I don't think you are getting, or acknowledging, LANL's
>requirements.
>
>
>Respectfully, I do understand, but an apology to you if you felt I was not
>properly acknowledging it. Our discussions on important topics such as
>these here should always be open and with honest challenges from all
>sides.
>No one is trying to skirt the issues, but rather face the issues head on
>together to improve on the working computing environments for everyone -
>as
>much as possible.
>
>That said, please allow me to address your most recent comments here. I
>know people must be getting tired of hearing from me on this today, so
>I'll
>try to do the best I can here and then allow folks to move on to more
>pressing communication and other topics.
>
>This has been good to discuss and I hope folks do not shy away from doing
>so
>in the future.
>
>
> Working with individual app developers is not feasible.
>
>
> We are not saying this is the ideal situation or one that scales for
>a single agency to work with every developer offering Apps, but are
>speaking
>about the best way to attack this particular problem given the need to
>mitigate these risks and what is available today. I will say that many of
>the App developers that are providing solutions leveraging the Enhanced
>Data
>Protection did so solely because they were approached by large customers
>with specific needs that the developer was willing / able to meet. This
>lead those developers to then incorporate those capabilities into their
>Commercial App Store offerings. It all starts with the developers. As
>more
>incorporate this into their Apps the better it is for you and every other
>customer - regardless of what the vertical market may be.
>
>
> For example, I want notes and calendar encrypted. Please
>tell me who at Apple I speak with to make that happen? I then have to
>contact Google about Drive. Then I have to talk to DropBox, Microsoft, and
>potentially hundreds of other developers to get them to flip a bit on
>their
>app? That is not feasible.
>
>
> Any need a customer has of any platform ultimately results in the
>customer communicating that to the provider. I hope you would agree that
>has been what you have been doing all along with your desktop/laptop
>systems
>as well over the years -- with Apple, Microsoft, Google, etc. Effective
>communication by customers and solid integration of those selected
>capabilities by vendors is what makes products better equipped to meet the
>needs of the individuals involved. Coming up with innovative ways to both
>solve the problems in a meaningful way and advance the platform is
>nirvana.
>
>
>
> When I say I want a global/system solution to turn it on, I
>mean int he System App. I wan enhanced data protection for all apps, all
>the
>time. Yes, I want the user to have to enter a password to do anything on
>the
>device except make an emergency call (or a regular call by entering the
>number directly).
>
>
>
>
> I have refrained from bringing up BB, but I will at this
>time. Their entire phone is FIPS certified.
>
>
> Folks who know me, expected me to stop here... :-) Respectfully, A
>Phone is not FIPS Certified. The cryptographic module used by
>Applications
>and Services on the device has been FIPS 140-2 Validated - Absolutely!
>The
>unfortunately thing about FIPS 140-2 is that it does not involve the
>review,
>validation or touch at all on ANY of the Processes, Applications or
>Devices
>that USE that cryptography outside of the boundary of the cryptographic
>module. To say it provides anything more is unfortunately not true. As I
>was noting the other day, I product could actually use a FIPS 140-2
>Validated Module with a NON-Approve Algorithm - not FIPS 140-2 Compliant
>at
>that point. I hope people really would read more about what FIPS 140-2
>Validation Promises and what it doesn't promise. There is so much
>misinformation about it throughout the community. For most it comes down
>to
>a checkmark and once they see a checkmark, it seems to take on a life of
>its
>own.
>
> I'll step down from my rant here... :-)
>
>
>
> Everything on the phone is encrypted/protected until the
>user authenticates (that is what I mean by "pre-use" authentication).
>
>
> If the phone is on, but the user is not logged in, I cannot
>access any _user_ data on the device. That is what I would like to see for
>the iPhone. When the phone is on, but the suer is not logged in, I cannot
>access any user data on the device.
>
>
>
>
> Right now I can copy a file (using itunes) to an app's
>sandbox on the iPhone/iPad.
>
>
> That is possible with iTunes only for Applications that have
>intentionally assigned the ability to share files via itunes. That is a
>specific API for developers to leverage. It was very necessary back in
>the
>days where iTunes was at the heart of all Comms to the iPhone/iPad/iPod
>touch. Since iOS 5 broke free from the requirements of being tethered to
>a
>computer, value add and use of that has sloped off significantly.
>
>
> I can then plug the device into any Mac, use iExplorer (or
>similar tool) and copy that file to the computer. Yes, _if_ the developer
>had implemented a class key above 7 (I think) that data would be safe, but
>since App developers generally don't do that, we have a problem.
>
>
> I am not sure what the reference is to Class Key above 7, but I
>think you mean to say the Developers need to protect sensitive files in
>Class A - "ProtectionComplete". With that I couldn't agree more! We are
>on
>the same page with that.
>
> All customers such as yourself have much more power than maybe you
>are giving yourself credit for. All developers, solutions providers, etc.
>need paying customers. Customers vote with their money.
>
>
> Yes, the iOS device _can_ be safe, but that is not good
>enough for government CUI data. We need for the device to always have the
>data encrypted, protected, and unavailable unless the user is logged in.
>hence the need for an App like Good, which degrades the user experience
>considerably.
>
>
> A point of clarification here. Users of an iOS device do not "log"
>into the device, but rather "unlock" the device. I am not picking on the
>way you said it here, but using this opportunity to point this out to many
>who do not know or understand that point. Each time a user "unlocks" the
>device, they are providing iOS with a non-stored value which is used as
>part
>of the tangled key for unlocking various Class keys. Prior to that
>"unlock"
>(while the device is running but is locked) those higher level Class Keys
>are unavailable to iOS and hence the Files are unable to be decrypted in
>any
>way by the OS or otherwise.
>
> The very good point that you have been making all along is that it
>only has this great protection of the files, if the developer properly
>utilizes the iOS provided services. I couldn't agree more with you on
>this.
>However, forcing ALL Applications and their data into "ProtectionComplete"
>now destroys the possibility of the devices to enable things like receive
>push mail, notifications (local or remote), alarms, MDM, etc. until the
>user
>would unlock the device and launch that specific application. It is not a
>usable scenario for the vast majority of users around the world including
>those in the federal government.
>
>
>
> I don't know what magic BB does to allow pushed e-mail, but
>I assume whatever they are doing it passed muster with the Feds.
>
>
> Assumption or verification ? :-) Remember, it comes down to risk
>management.
>
>
> I hope you understand what I'm requesting. Please let me
>know if there is a way to implement this, or if Apple has a solution that
>will be coming out.
>
>
> Apple is constantly working on raising the bar of data protection,
>despite users and developers :-), but we will not destroy the user
>experience. Some may choose to select third-party solutions on top of
>what
>iOS provides to achieve a level of assurance and risk they feel is
>acceptable. Many of those solutions do negatively impact the user
>experience, but may be a solution chosen by certain organizations. There
>will always be value for some in using a value added solution by a
>third-party. Everyone can't get everything they need from a single
>entity -
>it would be nice, but is not realistic.
>
>
>
>With respect to your other message sent later...
>
> From: "Marcus, Allan B" <email@hidden>
>
> Subject: Re: [Fed-Talk] CoreCrypto / CoreCrypto Kernel now
>in "In - Review"(CMVP)
>
> Date: February 20, 2013 4:52:13 PM EST
>
> To: "email@hidden" <email@hidden>
>
>
>
>
> I believe Apple has solved this issue with the keychains.
>
>
>
>
> As for the chicken and egg, it is irrelevant which came
>first as long as KFC is last! :-)
>
>
>I guess I am not understanding the context of this to be able to properly
>respond without potentially going off on the wrong path. I am always open
>to being educated :-)
>
>
>
>- Shawn
>________________________________________
>Shawn Geddis T (703) 264-5103
>Security Consulting Engineer C (703) 623-9329
>Apple Enterprise Division email@hidden
>
>11921 Freedom Drive, Suite 600, Reston VA 20190-5634
>
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Fed-talk mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden