Evolution Of Dedupe

Evolution of Data Deduplication

1

Evolution of Data Deduplication

(c) Druva Software 2010 February 11

Druva inSync – Overview and Advantage

What is Deduplication ?

• Specialized data compression technique

• Eliminates coarse grained redundant data

February 11

2

(c) Druva Software 2010

• Eliminates coarse grained redundant data

• Improves storage (and bandwidth) utilization.


Why is it so Important ?

February 11

3


8000

12000

16000

Ax

is T

itle

PC Data Vs Available Bandwidth*

Data (MB)

0

4000

8000

2000 2002 2004 2006 2008 2010

Ax

is T

itle

Data (MB)

Bandwidth (KB/Sec)

• Data doubling every 18 months

• Bandwidth grown only 10X last 10 years

• Data duplication

▫ 80% across PCs

▫ 45% across servers


Data Reduction at Target

February 11

4


• Duplicates removed at secondary storage to save

space


Synchronous/Inline Target Data Reduction

February 11

5


• Real-time/Synchronous duplicates removal

• Improved storage capacity management

• Slower than asynchronous deduplication


Source Based Data Deduplication

February 11

6


• Agent based deduplication

• Duplicates identified at source and removed from backup

• Saves bandwidth in addition to storage


Global Source Based Data Deduplication

February 11

7


• Duplicates compared against all sources

• Big Leap in bandwidth and storage saving


Granularity: Block Based Deduplication

February 11

8


• Granular block based comparison = Better Deduplication

• Works well across simple application and similar data streams

• Does not give very accurate results for complex apps e.g. Outlook or Exchange


Application Aware Deduplication

February 11

9


• Deduplicate logical information within data sets

• Works across applications

• 35-50% Better accuracy and storage/bandwidth savings


Granularity: Extent of deduplication

Fixed Blocks

• Good for single application

environments

February 11

10


Variable Length Blocks

• Block size determined by

heuristics/ rolling-

checksum

App-Aware Block Length

• Block size determined by

application data-structure

• Excellent for complex checksum

• Good for multiple data

streams

• Excellent for complex

applications like

Outlook, Exchange

• Delivers dedupe across

applications

1: 3X 1: 8X 1: 15X


Druva: App-Aware Deduplication

• Source Based

• Inline

• Global

February 11(c) Druva Software 2010

11

• App-Aware block sizes

• Supported applications –

▫ MS Outlook 2003/07/10

▫ MS Office 2003/2007/10

▫ PDFs

▫ JPEGs


Druva: Blackbird Storage Engine

• Near CDP ▫ All incremental backups

▫ Instant search based restores

• High Performance

February 11

12


(in memory) Hyper Cache

Blackbird storage engine

• High Performance� Distributed Caching

� Hyper-Cache (coming soon)

� SSD Support

• Scalable▫ Based on embedded Oracle DB

▫ 16TB, 200 parallel backups

• Simple� Software only solution

� Simple 20 Mins deployment

� Zero Maintenance

CDP + Dedupe File-system

(in memory) Hyper Cache

Oracle DB for Oracle DB for

meta-data

DiskSSD


How Does Druva Compared to Others ?

Source Based DedupeSource Based Dedupe

February 11

13


EMC, Acronis, IronMountain, Druva

, Veritas

Druva , Veritas, EMC, CA, Comvault,

IronMountain, Acronis, Atempo

Global DeduplicationGlobal Deduplication

InlineInline

Sub-FileSub-File

App-AwareApp-

Aware

, Veritas

EMC, IronMountain, Druva, Veritas

EMC Avamar, Druva inSync

EMC avamar, Druva inSync

Druva inSync


Why Druva

• A Fresh Approach Towards Backup

▫ Unique Cutting-Edge Technology

▫ Simplified Management

Everything is there. I checked

already… I've been in this game a long

time and I can honestly say that I have

never seen something so simple.

”▫ Enterprise Grade Support

▫ Affordable Solutions

• 600 Customers across 26 Countries

This, frankly, is brilliant. :-)

Cheers !

Christian R., Bechtel Corporation

Evolution Of Dedupe

Documents