by Hugh Emberson, CTO, StorReduce
Designing and deploying cloud-enabled backup & recovery solutions often leads to opportunities for reducing storage requirements and increasing efficiencies. Having effective cloud-native deduplication capabilities as part of your backup & recovery strategy can optimize migration, decrease the need for purpose built backup appliances like Data Domains, large tape archives, and enable cost reductions of up to 95%. In this session, StorReduce will provide best practices around data deduplication in relation to designing and deploying solutions around backup, archive, and general unstructured file data. They will also demonstrate how using a cloud native interface with scale-out deduplication enables generic cloud services like search inside all backups moved to cloud. They will guide the audience through two customer use cases from the financial services and healthcare industries.
2. § Massively Scalable Deduplication Software with a single namespace
Enables Primary and Secondary Backups to AWS and on the AWS
Cloud.
3. Problem: Legacy Storage for Backups
Server
Database
Backup
Appliance, e.g.
Data Domain
Tape
Primary Backups
• Daily backups retained for 30-90 days
• Hardware appliances
• Requires 2 appliances for redundancy
• Repurchase hardware every 3 years
Secondary Backups + Archival data
• Weekly, monthly, yearly backups,
retained from 1 month to forever
• Requires human time, hardware
• Not durable, hard and slow to recover
plus
Expensive hardware silos. Single points of failure and potential data loss. No scalability.
Limited durability. No “insight” into data using cloud services.
Data
Centre
4. Solution: AWS Cloud + StorReduce
Backups
Server
Database
Private or
Public
Cloud
97% reduction in bandwidth
and storage with dedupe
Data
Centre
+ Use cloud
services like
Search, AI, etc
5. Cloud
StorReduce
StorReduce
Purpose Built
Backup
Appliances:
e.g.Data Domain
Tape
Object Clone
Save 60%
Near instantaneous, unlimited copying of
an entire data set at PB scale, for free
Scale-out software, single namespace:
• Deduplication
• Replication between Clouds & Regions
• Cloning
Save up to 80%: Install on-premises in your
existing backup software to send backups and
archives to cloud. Remove PBAAs (like Data
Domain) and tape. Can also install on cloud.
Use any cloud service on backup data moved to
cloud.
6. Use Cloud Services on Backup Data Moved to Cloud
Cloud
Extract Metadata
& Data
Dark Data
Intelligence
E.g. Search inside all files for any
word or phrase for rather than look
through backup catalog document titles
Search,
Machine
Learning,
Analytics…
7. How StorReduce Works
S3
Interface
REST API
Admin Interface
Backup
Software
File
System
Gateway
Cloud
Services
Cloud Storage
Amazon S3,
HGST Active Scale,
Cloudian, HDS HCP,
Swift Stack,
IBM COS, ECS etc
S3
Interface
REST API
StorReduce Server
SSD
Index Data
Raw
Data
Deduplicated
Data
S3
client
S3
client
S3
client
VM, RPM, Docker,
AMI…
On cloud and / or on
premises
On-premises - requires only 0.2% of the volume of data being written to cloud in commodity SSD.
8. Characteristics
Fast
2500MBps / Instance,
Clustering to hundreds of
instances with scale-out.
Inline & Multi-Threaded
Variable length block
size av. 64KB
Stateless
All Data & Metadata stored in
private /Public Cloud
Software Defined
Deploy as VM, Docker
Container or RPM,
on-premises &/or on-cloud
Secure
Client Side Encryption
or S3 SSE
KMS and KMIPS compliant
Enterprise Ready
High Availability, Read
Replicas, & Cloud
Replication
Scalable
Unlimited Instances, no charge per
instance.
Scale-out version- single name-space
9. Architecture
Object Storage
Applications (Backups, Big Data, Unstructured Data …)
Read / Write at
LAN Speed
10s – 100s Gbit/s
Single Namespace
Highly Available
Single Dedupe Pool
24/7 throughput
No buffering
Commodity H/W
LAN Speed
10. Performance
9 server cluster, scales to many more clusters to increase throughput
• Ingest: 750TB per day, inline dedupe, sustained at 8.2 GiB/s
• Recovery: 4.5 GiB/s
• Deduplication ratio: 95%.
11. Backup to Cloud: Replace Backup Appliances and Tape
Cloud
StorReduce
Backupapplication
Ingest faster
Recover faster
Deduplication ratio = up to 95% reduction
Backup
Appliance
StorReduce
TCO Calculator
from AWS and on StorReduce website
Save up to 80%
StorReduce + Cloud:
1. No single point of failure
2. No silos
3. Scalable
4. Use cloud services on migrated
backups
StorReduce
Tape
StorReduce
Use cloud
services like
Search inside
all backups
Recover to on-premises or in
AWS Cloud and at multiple sites
12. Deduplicate Backups on Cloud: Enables lift & shift of all
Systems to AWS
Cloud
StorReduce
Backupapplication
Save up to 80%
StorReduce + Cloud:
1. No single point of failure
2. No silos
3. Scalable
4. Use cloud services on migrated
backups
StorReduce
StorReduce
Use cloud
services like
Search inside
all backups
Recover to in AWS Cloud – multiple Regions
Can also deduplicate Hadoop backups
13. Enterprises Keep their Existing Backup Applications
StorReduce is in NetBackup
Cloud Connector, and can send
deduplicated data to any cloud.
StorReduce enables NetBackup’s
OST features: Accelerator,
Accelerator for VMWare and
Optimized Synthetic Backups.
& any backup
system with a
S3 interface
Backup systems with a CIFS/ NFS interface, StorReduce is fully tested
with and uses QStar
NetBackup
14. § Require excessive hardware
§ Ruin the TCO to move to object storage
§ Risk data loss
§ Create silos
§ (Some) require you to removal all existing backup
applications and replace with theirs and rehydrate and land
on expensive disk to migrate data off backup hardware to
cloud
Other Migration Systems
15. Legacy Gateways: Data Domain Cloud Tier, NetApp AltaVault
Backupapplications:
NetBackup,Avamar,
OracleRMAN,…
Object
Store
Data Domain
w/ Cloud Tier
Data Domain
w/ Cloud Tier
1/3 1/3
1/3 1/3
2/3 2/3
2/3 2/3
Data Domain Cloud Tier (DDCT): 1/3rd of the data must
stay on Data Domain only 2/3rds can pass to object storage:
Tape
Data Domain Cloud Tier + Cloud:
1. Single point of failure – can lose data
2. Siloed
3. Expensive hardware cycle
4. Compromised geo-redundancy
5. Not a cloud interface (it’s CIFS/ NFS)
so no: search, data mining, cloning of
the data
Primary & secondary backups + long term archives
Data Domain
+
16. Equinix Professional Services (EPS): US
Healthcare Company
EPS contact:
ChristopherScalgione,AssociatePrincipal
EQUINIXPROFESSIONALSERVICES,CLOUD
1540Broadway#3900,NewYork,NY10036
Ecscalgione@equinix.com|M+17324069469
StorReduce and EPS Partner to Reduce a US Healthcare Company's on-cloud
Backup Costs by over 85%
Enables a Cost Effective Move of their IT Infrastructure intoAWS Cloud.
17. Equinix & Equinix Professional Services (EPS)
Solutions for infrastructure, connectivity, and cloud platform
PLATFORM EQUINIX
• Hybrid cloud enabling interconnection
platform with unprecedented global
reach.
CONNECTIVITY SOLUTIONS
1. Direct Connect Strategy Workshop
2. ECX & AWS Enablement
3. Network Transformation/Optimization
Direct Connect
Equinix Cloud Exchange
CLOUD PLATFORM SOLUTIONS
• Focus on Cloud Foundation & Migration
NETWORK
INFRASTRUCTURE
Equinix
Professional
Services
18. Customer’sGoal
Customer was backing up their organization’s servers to an on-premises purpose
built backup appliance, protecting approximately 10 Terabytes of backup data.
They required retention of daily backups for 180 days, which would total 1.8
Petabytes of storage.
Customer wished to move all of their on-premises IT infrastructure toAWS Cloud
for: increased durability for their data, scalability and the cloud’s superior disaster
recovery.
19. The Challenges
§ Frequent backups would end up with a large volume of data on-cloud
§ Cost of storing in cloud without deduplication was too expensive
§ Scalability - wanted a solution that scales with data growth
§ Didn't want to implement an expensive backup solution on-cloud
20. The Proposed Solution
§ Discovered StorReduce from the AWS Marketplace.
§ Projecting the costs of the customer storing straight to AWS S3 versus
using StorReduce to deduplicate and store on AWS S3 (including
StorReduce’s fees), Equinix found that StorReduce would reduce the
cost of storing their customer’s backups from $277,000 p.a. down to
$30,000 p.a., saving them $247,000 per annum, 89% of their
storage cost.
21. Solution Validation & Outcome
§ The POC was quick, effortless and successful.
§ The significant reduction in cloud storage cost for EPS’s client made it
worthwhile cost-wise to lift their IT infrastructure fully into the AWS
Cloud and to do their primary backups on-cloud rather than on-
premises.
§ EPS implemented CloudBerry Backup and went into production on
AWS using StorReduce deduplication.
§ EPS’s client has been in production on AWS cloud with StorReduce
since January 2016 very successfully..
22. Datacom: Insurance Company
Datacom contact:
AlexKennedy
NationalCloudManager|CloudServices
Email:Alex.Kennedy@datacom.com.au|Web:http://www.datacom.com.au
Ph:+61280943666| Mob:+61450125693
StorReduce and Datacom Partner to Move Insurance Company’s data off tape to
AWS Cloud
23. Datacom
Datacom is a leading VAR inAustralian and New Zealand and specialize in
architecting a cloud first future for clients
Customer’sGoal
Customer was storing their secondary backups and archival data to tape. They
wanted to avoid the cost of repurchasing tape hardware and manual daily tape
rotations.
Customer wished to move all of their data off tape to store it toAWS Cloud for:
increased durability, scalability and cost savings.
24. The Challenges
§ Ensuring a cloud first future for the company’s data
§ Moving data from multiple sites to the AWS cloud over existing
bandwidth
§ Cost of storing in cloud without deduplication was too expensive
§ Workflow – wanted a solution that worked with their NetBackup
environment
25. The Outcome
§ StorReduce and Amazon Cloud were selected as the storage medium
for this company’s long term and secondary backups.
§ The client has been in production with Datacom, StorReduce and AWS
Cloud since December 2016 at Petabyte-scale.
26. The Outcome (…cont’d)
§ Instead of the cost of repurchasing tape hardware and manual daily
tape rotations, the entire process is fully automated and fail-safe:
Veritas NetBackup takes the backup, StorReduce deduplicates the
backups on-premises inline and transfers them securely with reduction
in bandwidth and reduction in backup window time, to store
deduplicated in the AWS Cloud.
§ The client now has the peace of mind that the backups are stored on
the highly secure, redundant AWS Cloud with no single point of data
loss, unlike legacy tape storage.