Storage Spaces Direct Data Deduplication and Azure File
























































































- Slides: 88
Storage Spaces Direct, Data Deduplication, and Azure File Sync best together
Storage Spaces Direct, Data Deduplication, and Azure File Sync best together or mix-and-match
Storage Spaces Direct is software-defined, shared-nothing storage.
Use commodity hardware to build lower-cost storage infrastructure
How does Storage Spaces Direct accomplish this?
How does Storage Spaces Direct accomplish this?
Cluster the servers
Create a storage pool
Create a new volume
C: Cluster. StorageVolume 1 Create a new volume
Resiliency on Storage Spaces Direct
Mirror Resiliency on Storage Spaces Direct
B’ B’’ B Mirror Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies Mirror Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies P 1 B Mirror P 2 Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ PD 34 CD Copies P 1 B Mirror P 2 Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ C PD 34 Copies P 1 B Mirror P 2 Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 Parity Bits Data P 1 B Mirror P 3 P 2 Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 Parity Bits Data P 1 B Mirror P 3 Parity P 2 Mirror-accelerated Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 AB’’ AB Parity Bits Data P 1 B Mirror P 3 Parity P 2 Mirror-accelerated Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 P 1 Parity P 12 B AB Data Parity Bits Data B Mirror P 3 AB’’ Copies P 2 Mirror-accelerated Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 Parity Bits Data P 1 B Mirror P 3 Parity P 2 P 12 B AB Mirror-accelerated Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 Parity Bits Data P 1 B Mirror P 3 Parity P 2 A PB 12 Mirror-accelerated Parity Resiliency on Storage Spaces Direct
B Data B’ B’’ Copies D C P 4 Parity Bits Data P 1 B Mirror P 3 Parity Data P 2 A Parity Bits B P 1 P 2 Mirror-accelerated Parity Resiliency on Storage Spaces Direct
Storage Spaces Direct and general purpose file server
Storage Spaces Direct and general purpose file server
Storage Spaces Direct and general purpose file server STORAGE POOL
Storage Spaces Direct and general purpose file server VOLUMES STORAGE POOL
Storage Spaces Direct and general purpose file server Run files in virtual machines Isolate file servers by function VOLUMES STORAGE POOL Higher availability, performance Storage Qo. S
Data Deduplication optimizes data storage by single-instancing common chunks of data
Save up to 95% on any hardware backend
How does Data Deduplication work?
How does Data Deduplication work?
A B C C A A C D D C Break files into variable-sized chunks
A B C C A A C D D C Identify unique chunks
A B A C C A C D D C Identify unique chunks
A B C C A A C D D C Identify unique chunks
B A C C A A C D D C Identify unique chunks
A B C C A A C D D C Identify unique chunks
A B C C A A C D D C Identify unique chunks
A B C C A A C D D C Identify unique chunks
A B C C A D A C D C Identify unique chunks
A B C C A A C D D C Identify unique chunks
A B C C A A C D D C Place chunks in the Dedup chunk store
A B C Place chunks in the Dedup chunk store D
A B C D Replace original file stream with a pointer
RP A B C D RP RP Replace original file stream with a pointer
RP A B C D RP RP Replace original file stream with a pointer
… Target workloads …
Data Deduplication availability
Azure File Sync file sync sy nc file sync file Fabian Uhse Branch Centralize file services in Azure storage Cache in multiple locations for fast, local performance PROGRAM MANAGER, AZURE STORAGEUtilize Branch HQ Cluster cloud-based backup and fast DR
Paa. S S Iaa RE SM ST B Work Folders West Europe East US File. Share File Share SMB NFS Branch Office Paris Azure Backup Vault Users Applications ! HQ Server Seattle Branch Office New York
Multi-site sync Cloud tiering Direct cloud access Integrated cloud backup Fast Disaster Recovery Keep a data set in sync across multiple locations Turn a server into a lightweight, disposable, performance cache for Azure Files Native file format with SMB and REST access Point-in-time recovery and retention policies without redundant data transfer Fast namespace sync with no need to wait for data restore
ETA: Q 1 CY 18 Ignite 2017 Dec 2016 Private Preview Multi-site sync Cloud tiering Rapid server DR Evaluation workloads Single region Public Preview Integrated backup Production workloads Few regions General Availability Cloud access with AAD DS Cross-region sync Public regions Full SLA
Storage Spaces Direct, Data Deduplication, and Azure File Sync
Storage Spaces Direct, Data Deduplication, and Azure File Sync
BRK 2286, Microsoft Azure File Sync seamlessly extend file services across servers and cloud, Tue 10: 45 - 12: 00 BRK 2158, Windows Server Fall Release technical foundation, Wed 10: 45 – 12: 00 BRK 2372, Azure Disks: Scale, performance, and ease of use, Thu 10: 45 - 12: 00 BRK 2288, Azure Storage: Digital transformation in key industries, Thu 12: 30 1: 45 BRK 2278, What’s new in Windows Server, version 1709 for the software-defined datacenter, Thu 2: 15 -3: 30
https: //myignite. microsoft. com/evaluations https: //aka. ms/ignite. mobileapp
C: Cluster. StorageVolume 1 Create a new volume
What’s so interesting about this volume? C: Cluster. StorageVolume 1
What’s so interesting about this volume? Re. FS C: Cluster. StorageVolume 1
What’s so interesting about this volume? Performance Re. FS Tiered Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Mixed resiliency Capacity
Mirror P 1 P 2 Parity Performance Capacity
What’s so interesting about this volume? Performance Re. FS Tiered Mixed resiliency Capacity
Data Deduplication is now available on Re. FS
Performance Tiering-aware deduplication Capacity
Performance Tiering-aware deduplication Capacity
Performance Tiering-aware deduplication Capacity
Performance Tiering-aware deduplication Capacity
Tiering-aware deduplication Performance Dedup Chunk Store pinned to Capacity tier Capacity
Tiering-aware deduplication Performance Dedup Chunk Store pinned to Capacity tier Capacity
Tiering-aware deduplication Performance Dedup Chunk Store pinned to Capacity tier Capacity New optimization mode: Near Inline
Tiering-aware deduplication Performance Dedup Chunk Store pinned to Capacity tier Capacity New optimization mode: Near Inline
Tiering-aware deduplication Performance Dedup Chunk Store pinned to Capacity tier Capacity New optimization mode: Near Inline
Tiering-aware deduplication Performance Dedup Chunk Store pinned to Capacity tier Capacity New optimization mode: Near Inline