to dedup or not to dedup ? deduplication deep dive
TRANSCRIPT
About Me
Kamil Bączyk
• Core & BPOS Consultant \ Trainer • MCT, MCSE, MCSA, MCITP, MCTS, MCP, ITIL, CEHv7, etc ….
• @ Facebook, LinkedIn, Goldenline, Xing …• [email protected]
FactsBoiling Points:• File Data Growth
Data demand rises Increased stress – backup solutions
• Storage TCO Disk cost Data Management
• Consolidation• Data decentralization
Offline and secure data WAN / LAN Access IDC Worldwide File-Based Storage
Forecast, doc #231910, 2011
Facts – what is it ?
• Goal :use less storage • Method :Check and ensure that identical content in multiple files (big) is only stored once• Block-based, post-process,
transparent solution
Dedup
Facts – what is it ?MODELS:• Source :Prevent transferring data, if duplicate• Inline :Perform dedup when data is written (NTFS Compression, Slow• Post-Process :Works after, in the background, Used in Windows Server 2012
Dedup
Facts – what is it ?Other ways:• SIS (Single-Instance-Storage):File based solution in W2K• NTFS Compression :Inline, CPU based• NTFS hard links :Not transparent, file-based
Dedup
Facts – how it works ?
• Segment data into „chunks”• Identify duplicate chunks• Maintain a single copy of each
„chunk”• Compress• Reference existing files with
reparse points • Default (5day policy)• Free
Facts – how it works ?• Manage by PowerShell, WMI, Server Manager,• Transparent for data – Mini Filter• Maintain a single copy of each „chunk”• Compress Data Chunks• Dedicated Service (DDPSVC)• Default (5day policy)• Audit - Event Viewer\Applications and
Services Logs\Microsoft\Windows\Deduplication\
• Chunks files 32-64 KB• CRC Checking – Hot Spots
Facts – how it works ?
• Only on Windows Server 2012• Filter Driver• Windows Cache – Dedup Aware• No GPO• Doesn’t work on CSV but
supports cluster data • No support for Boot Drives or
System Drives
• Works on NTFS, no ReFS support• Excluded specified files by default
or manually• DDPSVC – checks 100Gb ~ 1 hour• Default policy 5days • Scheduled process• Live SQL DBs or Exchange • Live Production VMs / VDI
Discussion ?When to use it ?• VHD backup library• Folder redirection servers
• DB backup volumes• Software deployment shares
Thank You
Kamil Bączyk
• Core & BPOS Consultant \ Trainer • MCT, MCSE, MCSA, MCITP, MCTS, MCP, ITIL, CEHv7, etc ….
• @ Facebook, LinkedIn, Goldenline, Xing …• [email protected]