Open Stack

On Mon, Nov 4, 2013 at 9:12 AM, Chris Friesen
<chris.friesen at windriver.com> wrote:
> On 11/03/2013 08:39 PM, Qixiaozhen wrote:
>
>> In my opinion, we should rethink the way of wiping the data in the
>> volumes. Filling in the device with /dev/zero with “dd” command was the
>> most primitive method.  The standard scsi command “write same” could be
>> taken into considered.
>>
>> Once the LBA was provided and the command was sent to the SAN , the
>> storage device(SAN) could write the same-data into the LUN or volumes.
>> The “dd” operation  can be offloaded to the storage array to execute.
>
>
> This will offload the CPU on the compute nodes, but it still uses disk I/O.
>
> Personally I like the elegance of giving each user a certain amount of I/O
> and giving them the option of unsecure or secure deletion.  If they choose
> the secure deletion then that disk bandwidth comes out of their quota and
> the volume size is still billed against them until it is done being deleted.
Well, I have mixed feelings with this. On one hand, I agree, that its
a commodity and the costs should be passed on.

> By asking for the secure deletion they are essentially still making use of
> those resources until the deletion is complete, so I think it's fair that it
> count towards their quota/usage stats.
Here's where I have the problem. When the service was sold, were they
explicitly charged for the I/O to set up the VM? I expect not.

When the service was sold, were they told they needed to buy
additional compute time to ensure a information leak or data breach
does not occur? I expect not.

Perhaps the wipe should be included and amortized in the cost of the
account and setup. If wipe is not used, then the provider gets to
pocket the proceeds. That's how most certificate and public key
infrastructures are setup and the CAs aren't complaining (the cost is
in revocation, not issuance).

Jeff