Accelerator process flow 1 Net Backup Accelerator Challenge

Accelerator process flow 1

Net. Backup Accelerator • Challenge – Customers can’t complete their full backups within their backup window • Solution – Significantly reduce time required for full backups of file systems • Features – Net. Backup client reads, de-duplicates and transfers only changes to media server and sends only metadata for changes to Net. Backup catalog – Optimized synthetic full backup generated – NTFS change journal utilized to remove need to enumerate file system – Accelerator functionality supported with all client OS platforms (dedupe is performed on client, media server or appliance) 2 2

Net. Backup Accelerator • Realized Value – Full backups to Net. Backup dedupe, Cloud or OSTvendor storage pools performed in similar amount of time as incremental backups – Reduced CPU and I/O overhead on Net. Backup client 3

Accelerator – Overview • Significant time reduction in performing full backups to file systems and VMware • Align with existing deployments – – Leverage existing architectural framework & infrastructure Allow use of existing data protection management paradigm Seamless integration with existing Net. Backup environment. Evolutionary approach to providing a new solution • Reduce catalog data for full backups sent to Master Server • Improve RTO and RPO • Requires storage supporting Optimized Synthetics and Optimized Duplication, or Cloud 4 4

Traditional full backup 1 • Read all data 2 • Send all data Master Server Client NBU Catalog 4 Application Media Server bpbkar 2 bpbrm 1 bptm File System 3 Disk Storage Unit 3 • Process all data 4 • Send all metadata to catalog Results: • High I/O cost and moderate CPU cost of long duration on client • High network cost • High storage cost and catalog growth 5

Media server deduplication backup 1 • Read all data 2 • Send data to dedupe Master Server Client Image Catalog 5 Application Media Server bpbkar 1 File System bpbrm 2 Dedupe Engine 3 bptm 4 Dedupe Storage Unit engine 3 • Dedupe data 4 • Write data to dedupe pool 5 • Send metadata to master server Results: • High I/O cost on client and high CPU cost on the media server • Catalog growth • Minimized storage cost 6

Net. Backup Accelerator backup using deduplication Client 1 • Read client track log 2 • Send changed data to Master Server NBU Catalog 5 Application 1 Client track log bpbrm 2 Dedupe Engine 3 bptm Synthesis Engine bpbkar Media Server 4 NTFS Change Journal media server 3 • Dedupe changed data 4 • Write to dedupe pool 5 • Send metadata to master server Results: • Quick backups • Low CPU cost • Minimized network bandwidth and storage cost Dedupe Storage Unit 7

Process flow: full backups after initial backup Enable Change Journal Enable Accelerator Client side Dedupe changes on client Transfer new segments to media server Send changes to database Deduplication type? Media server Send changes to media server Dedupe changes on media server Store segments on dedupe disk Send full backup metadata to catalog Create optimized synthetic full 8

Thank you! 9