quote:

Originally posted by Rey HRH:

quote:

Originally posted by sigmonkey:
Data can be sandboxed and separate clean data streams co-located on multiple volumes and or servers,
A very low cost and simple cloud solution is Spider Oak.

https://spideroak.com

Got it about backing it up. But I tried reading and re-reading what I quoted above several times and I can't make sense out of it. Can you please break it down or explain?..

Sorry. Did not mean to over speak the point.


Cloud backup that has your data backed up on different servers instead of a single server (and oftentimes virtualized server instances) and/or different volumes (and different drives or media servers) and isolated from any connection oar applications except for strict access that is not a continual connection state, but only accessed for the adding of new data, and fetching of stored data.

Separate data streams containing the data, and or pieces of that data in a method that provides more than one instance of the handling of the data.

As an analogy, having a storage facility that is robust, secured and only allows specific verified and authorized "trucks" bringing in the "stuff" to be stored, while making copies at eh point of origin (your place) and then transporting the multiple copies in parts to more than one location and then further storing the copies in multiple units in the multiple locations.

It is easy to do with data, to make all these copies and use the procedures to store in multiple locations, and only add the changes.

Say you have a couch, made of wood, metal and fabric. Mathematically this couch and the new couch you buy, or if you recover the couch, are very similar from a molecular standpoint, and that "math" can be used to only add the differences from the "molecular math" to then contain the information to recreate any iteration of the couch, to include every iteration back to "your place" from the storage place, without them having to store a full copy of each couch.

In effect, the "recipe for couch" and you can make any couch in that recipe from 1 and 0.


This is "loosely" how data de-duplication, and also how delta differencing works to only back up the block in data that have changed, and only one instance of data stored, no matter how many times you use the content of that data in different files.

I have 26 servers in 4 states with a total of 90 TB of data that is reduced to about 5TB of stored backup data, and able to grab a single file to everything, from any point in time, and the methodology ensures no "contamination", as the backups are "scattered" to multiple places, and are "clean" because of how the data is created, accessed and retrieved.

And no processes access it in a manner that permit it to be "contaminated" by ransomware and the like, (as well as endpoint, and back end process hardening in our network environment).

It's a might big back of onions, with lots of layers.

But I keep my hammer and saw handy in case it ever all goes to shit. I can bang nails and dig ditches to by a box of Raman noodles...