I'm putting this up for your thoughts, on my thoughts....and trying to tune others in

FWIW RE: ZeroSnap

Some times I get things on my mind and they just will not leave until I can come up with a conclusion to my original consideration.

article

video VMware

First off, you have to consider what a VM (virtual machine) is. It's the environment that functions between the storage memory(harddisk) and the working memory(RAM) and the methods utilized to track data that are stored on the hard drive and preserve the metadata records. Simply, a file(metadata) is created by an application(operating system) that points to all the data stored. This file boots to RAM in part, or in full when the computer is turned on. Because this file resides in RAM it is considered to be in virtual state and is venerable, as RAM is a volatile memory, when there is power interruption you lose that Virtual file and everything on it.

Now, if you paid attention to the video, it was explained in dissertation 1, legacy hard drive, exactly what I explained above. Legacy methods in cloning the full volume, that being data and virtual files(metadata)....basically copying the hard dive contents and re-writing a new volume on the same hard drive. Copying up to the RAM and re-writing back down to the drive, if you have a desire to clone the whole volume. For regular operation of managing the virtual files(metadata) you do not clone the whole volume, the metadata files are updated when user data changes, managed in a substantial RAM resource as metadata files are large..

With that, dissertation 2, their idea of CopyX and SSD....explains how they perform the operation of cloning, with what appears, without utilizing RAM resource. They simply copy (A) array blocks of the flash....and re-write the contents to (B) array blocks... all the data and metadata(virtual files). No substantial RAM resource utilized other than to create a file(metadata) of where this new volume is located on the SSD. Thing is, that explanation is misleading, it would require RAM, depending on the flash makeup of the SSD, ....NAND or NOR. For NAND it requires RAM, for NOR it does not, however there are cons for NOR used as storage memory.....SSD's will be NAND.

Copy, in general, requires RAM no matter what the storage media is. You can not move data from one place to another without it. e.Digitals trick is in how they size the RAM and that is because of the operating system.

For this, I have a concern of how they are moving the data without distinguishing a RAM use, as they do in dissertation 1.

For dissertation 3 ZeroSnap. With this, they do not copy the whole volume, they utilize the original data stored and there is no re-writing of it. For this, they create what they define as a compressed copy of the metadata(the virtual directive) and call it a new volume, though it's connected to the original data..

Now lets look at this compressed Metadata. In this ZeroSnap process the metadata is seriously compressed still matching the general data address locations, IMO, they are re writing the metadata into a logical overlay as e.Digital does. The general legacy data does not have embedded operating system directive, however, the metadata is re written into a header scheme and contains the operating system directive, there by eliminating the over head to maintain the metadata. Maintaining in this fashion they do not need to utilize substantial RAM resource. With this they can combine legacy data and newly created data seamlessly if they wish. To do this it would require an operating system that would write data under header schemem....and match the RAM to the storage memory write, or however they wish ......their offering has the ability of "all I/O sizes"

Keep in mind, e.Digital owns patent rights(445), as the first to devise an operating system that implements header schemes. Though header issues are not per say novel, the way that the MOS is orchestrated in writing data to storage memory...it is novel, and the memory type does not matter.

doni