openio serverless storage
TRANSCRIPT
ServerLess Storage
most
88 % and growing fastCapacity
Storage market evolution (80/20 rule)
Latency sensitive
IOPS
High Capacity
High frequency modifications
Low frequency modifications & immutable data
High frequency modifications
most
88 % and growing fastCapacity
Low frequency modifications & immutable data
Storage market evolution (80/20 rule)
Latency sensitive
IOPS
High Capacity
What is driving up capacity?
What is driving up capacity?
What will drive up capacity?
What will drive up capacity?
Redefining Big Data
Redefining IT InfrastructuresRedefining IT Infrastructures
Next Generation ApplicationsNext Generation Applications
Creating metadata from data
Ingest Search
Store
Analyse
Add metadata
Ingest Search
Store
Full Text Index
Spam sample long-term archiving and search
Real life use cases
Computing models, Evolution
Containers
Not storage centric
Virtualization
Complex and expensive
ServerLess
Event-driven
X86, one of many options now
X86
General purpose
Coprocessors
Specialized applications
ARM
Efficiency, IoT and mobile
All-flashStorage
Specialized HCI
Hyperconvergence, all but efficient
General purpose
HCI
In-memory Storage
Lambda-like Enabled storage
Capacity-driven
CPU
Latency-sensitive
RAM
TYPE OF WORKLOAD
COMPUTE RESOURCES
Storage and infrastructure efficiency
Run your application on the storage infrastructure, where data resides!
Any media
Any hardware
OpenIO SDS
GENERATEAPIs
INGEST
PROCESS
SERVE
STORE
REPLICATE
TIERING
Run your applications where data resides!
All-flashStorage
Specialized HCI
General purpose
HCI
In-memory Storage
Lambda-like Enabled storage
Capacity-driven
CPU
Latency-sensitive
RAM
TYPE OF WORKLOAD
COMPUTE RESOURCES
OpenIO
New object storage use cases sustainability
$/GB
Backup repositories CollaborationStorage
consolidation Big Data lakes and IoT
New object storage use cases sustainability
CollaborationBackup repositories
Big Data lakes and IoT
Storageconsolidation
Ease of use$/GB Performance Flexibility
Storage Array x86 Commodity Servers
EthernetTCP/IP drives
1995 - 2005 2006 - 2016 >
Large capacity storage evolution
Scale-up and RAID 60-90 slots x86 Commodity Servers
EthernetTCP/IP drives
• Impracticable • Large failure domain • Low performance • Inefficient
• Half-baked technology
Large capacity storage $/GB tradeoffs
OpenIO SLS-4U96 appliance
Open Source softwareCommodity hardware
# Reduced
cost and TCO
Object Storage
# Based on SDS
object storage platform
ARM-based nano-nodes High-performance backend
# Smallest failure domain
Massive I/O parallelization
OpenIO SLS, the scale-out storage infrastructure in-a-box
Scale-out radicalization1 disk per nano-node
Massive paralellization96 nano-nodes per chassis
No-compromise throughput40Gb/s front-end
Power efficientARM-based architecture
Small footprint96 3.5” HDDs in 4 rack units
No Single Point Of FailureRedundant links, PS and Fans
Efficient & No SPOF
Scalable & Powerful
SDS enabled
Easy to use
in a single appliance
OpenIO SLS
How
• Dual-core ARM-v8 CPU
• RAM, flash memory, 2 * 2.5gb/s Ethernet links
• 3W power consumption and HDDs Power management
• Supports 8,10,12 TB HDDs
Hyper Scalable Storage
Nano-node
• N+1 power supplies and cooling units
• Chassis Management
• 2x 6-port 40gb/s Ethernet switches for front-end and back-to-back expansion
• Up to 96 hot-swap nano-nodes
No Single Point of Failure
SLS 4U96
• N+1 power supplies and cooling units
• Chassis Management
• 2x 6-port 40gb/s Ethernet switches for front-end and back-to-back expansion
• Up to 96 hot-swap nano-nodes
SLS 4U96
40 Gb/s Ethernet switch
40 Gb/s Ethernet switch
6x ports
6x ports
Nano-node
HDD /SSD
2,5 Gb/s HS-SGMII
#1
#96
No Single Point of Failure
2,5 Gb/s HS-SGMII
• Standard Object APIs to leverage nativelythe platform: OpenIO REST/HTTP, Amazon S3 and OpenStack Swift
• Industry File-Sharing Protocols: NFS, SMB, AFP and FTP
• Several data protection schemes and cluster topologies
• Ease of Use. GUI, APIs, CLI
• Lightweight backend design
Same software, same capabilities
SDS
The Object Storage Appliance for all kinds of organizations
3.2.1.
Minimum configuration starts at 12 HDDs, up to 96 in a single chassis for a total of 1152TB. Multiple chassis supported in a single cluster.
EfficientEasy to use
4.
All the features available with SDS.
Scale-out cluster in-a-box. Easy to deploy, manage and use.
Low power and datacenter footprint but with high throughput. Expandable by one disk at a time and No SPOF.
Scalable Feature rich
ServerLess Storage