Monday, December 9, 2019

Why new Data Domain models? - DD6900 / DD9400 / DD9900 backup/recovery appliances

In September 2019 Dell announced 3 new Data Domain models (DD6900 / DD9400 / DD9900)
Shall we take under considerations new Data Domain models?

If so why?
What is the architecture?



Below is the summary of changes in new Data Domain models:
  • DD6900
  • DD9400
  • DD9900

Automated protection against ransomware / hacker.
Any backup performed on new Data Domain models is automatically protected for defined period.
Neither ransomware nor hacker can delete it.
How is it possible?
https://youtu.be/xx-xZC9JtoM?t=1169

Reducing disk space required for backups
New models require 10-30% less space for storing backups comparing to previous ones.
Wow! Can it be?
https://youtu.be/xx-xZC9JtoM?t=191

Starting even 64 Virtual Machines from new Data Domains with 60 000 IOps performance!
Hugh, can backup be faster than production?
https://youtu.be/xx-xZC9JtoM?t=961

Faster restores from new Data Domain

Capacity on demand
In new Data Domain models we can get more space but pay for the one that we use.
Details:

Faster internal components
Does it matter?

Scalability from 48 TB to 1 250 TB
Details:

Networking
100Gb Ethernet card is possible!

Very little space required Can you store 1.25PB in single rack?
https://youtu.be/xx-xZC9JtoM?t=1512

Monitoring and visualizing online all its components
Nice…

Upgrades
How can it move to new models?

--------------
New Data Domain models have a number new & exciting things.
It makes sense to have a closer look on them...

Presentation about new Data Domain models (the one used in video):
http://gurago.pl/s1/NewDataDomains_Detailed_DD6900_DD9400_DD9900.pdf

Only immediate recoveries!
BackupRecoveryGuy

Wednesday, August 21, 2019

How to keep long term backups? - Tiering or... something simple!


Recently, customer asked me the question:
      I’m looking for a discussion document which outlines best practice for tiering backups.
This question comes for ages.

Why?
Idea of tiering in backup comes from our internal dream:
Let’s have fast, secure and good media (storage) for recent backups/data
but at the same time
Let’s keep old backups as cheap as possible

Recent backups we usually keep on Data Domain (https://youtu.be/f4T5eyLz5Rs).
Data Domain gives us huge performance, security, and restore.
This is what we require from protection production: no load, fast backup and easy, fast restore.

For older backups copies (archive backups) we really do not know what to do?
We just need old backups to keep cheap...

We have couple of options for storing old backups:
        Tape
Object storage (local/cloud)
Or maybe…

Why not just keep 5 years backup on Data Domain?
Local Data Domain / Data Domain in cloud – as one wishes.

During one of the workshops I asked customer:
       How long do you keep backups on Data Domain, how long on tape?
He said:
       Daniel, I have no time and money to play with backup tiering.
       I have bought 2 Data Domains (little bigger) and I just replicate between 2 cities.
       I have 90 copies on each Data Domain (30 daily for last month and 60 monthly for last 5 years)
       No touch, same cost, dream life.

My experience (I would choose it for myself) is that little bigger DD - for both production backups and archive backups - in many cases is better than tiering.
Money are similar, simplicity is incomparable.

Of course, every case is different, but…
Most simple solutions are the best ones!

----------
And let me quote my great friend from UK:

When I started dealing with backup I was said that the more we retain on Data Domain, the better de-duplication is.
And that is so true…

The current push for tiering backups in reality makes more problems rather than solving any.
Tiering backup is:
  • Expensive
  • Complicated because of restore 
  • And many other compatibility issue
---------
Why keeping last backups on great media (like Data Domain) and moving older backups to tape/object provides complication, less flexibility, additional costs?

With tiering in backup, we have 2 different media, software, policies to manage those 2 different backup media?

Why do we do that all complication?
To make backup cheaper.

But… Even hard costs of tiering like additional storage (second tier) + potential licenses for tiering in many cases eat that whole difference in price between 2 different media.

Not mentioning soft costs like management, integration, know how, …
This is what my customer just said from his experience.

So having fantastic media like Data Domain, we can just increase a little bit its space – not much due to de-duplication – and we can enjoy f.e. 5 years retention!
With no problems, no management, easy.
And with similar or even less money…

And... Data Domain can be in the cloud using Object Storage as the space for backups.
But this something for another article...

----
Again:
Most simple solutions are the best ones!

Only successful recoveries….
BackupRecoveryGuy