Agenda
1
2
3
4
5
Introductions
Use	Cases
Question	&	Answers
Problems	with	Legacy	Deduplication	
Moving	Deduplication	to	a	Cloud	1st	Model
About Cloudian		
3
CLOUDIAN PARTNERS
TECHNOLOGY
GLOBAL PRESENCE
HQ:	San	Mateo,	CA;	
Offices	in	US,		Japan,	China,	EMEA
TOP TECHNICAL TALENT
With	deep	experience	in	storage,	
big	data	&	enterprise	software
TOP TIER INVESTORS
Intel,	Fidelity	,	Goldman	Sachs,	
INCJ	
BORN IN JAPAN
BASED IN SILICON VALLEY
To	fix	storage	problems	created	by	
exponentially	growing	data
2011 $44	M 80 Global
CHANNEL
About StorReduce
Firsts:
• First	cloud	centric	deduplication	engine
• 1 of	8 launch	partners	Google	Coldline
• 1 of	4 launch	partners	Amazon	S3-IA
Clients	on	3	Continents
HQ:	Sunnyvale,	CA
Offices:	USA	&	Asia	Pacific
“We	use	StorReduce	in	production	for	
a	major	US	healthcare	company.	
StorReduce	does	exactly	what	it	says	it	
will.	It's	market	changing.”
Bill	Young,	Director	Professional	
Services,	Equinix
Problems	with	
Legacy	Deduplication
Not	Built	for	Cloud
Legacy	Deduplication	built	to	support	CIFS/NFS
• Require	Expensive	Disk	Based	solutions	
• Data	is	locked	into	and	accessible	only	from	the	appliance
• Unable	to	handle	high	latency	connections
• Not	Stateless
• Makes	moving	to	the	Cloud	harder	&	more	expensive
Not	all	Deduplication	is	Equal
Fixed	Block	Sizes	limit	efficiency
Require	a	Scale-Up	approach
No	Global	Deduplication	across	Datasets
Databases	become	fragmented	on-recovery
Fragmented	data	increases	storage	costs
Poll
If	you	could	reuse	your	backup	data,	
how	would	you	use	it?
Data	Mining
Search
Dev/Test
Introducing	StorReduce with	Cloudian
A	Cloud	First	Approach	to	Deduplication
10
Cloud	Ready	Backups
11
Cloud	First	Deduplication
StorReduce deduplicates S3	streams	of	data
Perfect	for:
• Primary	Backup	Data
• Tape	Replacement
• Hybrid	Storage
• Database	Backups
• Hadoop	Backups
• Other	Unstructured	Data
• Variable	Block	Deduplication
• Reduce	Network	Bandwidth	
and	Data	Storage
• Remote	Office	Friendly
StorReduce is	…
Fast
1400MBps	/	Instance,
Inline	&	Multi-Threaded
Stateless
All	Data	&	MetaData stored	
in	Cloudian/Public	Cloud
Scalable
80PB/Instance
Unlimited	Instances
Software	Defined
Deploy	as	VM,	Docker	
Container	or	RPM
Secure
Client	Side	Encryption
or	S3	SSE
Enterprise	Ready
High	Availability,	Read	
Replicas	&	Cloud	Replication
Cloudian	Scale-Out
13
HyperStore:		Scale	Out	Performance	&	Capacity
Commodity	Servers Scale	Out Durable Simple	to	Use
Heterogeneous	Nodes
500TB+
Usable
Multi-Rack,	Multi-Datacenter,	Multi-Region
100TB+
Usable
Start	Small	&	Expand
3PB
Usable
Storage	without	Compromise
REPLICATION
(RF=1,2,3,4)
ERASURE	CODING
(N+1,2,3,4)
CONSISTENCY
(Strong	or	Eventual)
Policies	Applied	per	Bucket
• Choose	Durability
• Erasure	Coded
• Replicated
• Choose	Geographic	Availability
• One	Datacenter
• Multiple	Datacenters
• Replicated	Erasure	Codes
• Plus	+
• Strong	or	Eventual	Consistency
• Tier	to	Amazon	S3	or	Google	Cloud	Storage
• Encryption
Encryption
SSE	or	SSE-C
Geographic	Dispersion
Assign	Copies	per	DC
New	Economics	of	Deduplication
JBOD
StorReduce $0.07/GB
Cloudian	HyperStore	
$0.25
*Based	on	Analyst	reported	Street	Price	for	Object	Storage	and	Deduplication	Appliances	
JBOD
CPU/	Controller
Cloudian	&	StorReduce
$0.32/GB
Data	Domain	or
NetApp	AltaVault
$1.00/GB
• 66%	Cost	Savings	including	Hardware
• Pay	for	what	you	need
• No	Forklift	Upgrade	every	3-5	Years
Poll
What	is	your	current	Backup	Application(s)
NetBackup	
CommVault
IBM	Spectrum	Protect	(TSM)
Veeam
Oracle	RMAN
Use	Cases
A	Cloud	First	Approach	to	Deduplication
18
NetBackup
NetBackup	v7.7	Enables	S3	Backups
• Full	/	Incremental	Only
• V8.0	no	announced	support	for	Dedupe
Clients NetBackup
Media	Server(s)
StorReduce
Cloudian
HyperStore
19
Commvault
Clients Commvault
Media	Agent(s)
StorReduce
Cloudian
HyperStore
Deduplication	pool	becomes	fragmented
Garbage	collection	issues	on	object	storage
Questions?
Thank	You
www.cloudian.com
Cloud	Storage	for	Everyone

S3 Deduplication with StorReduce and Cloudian