Overview 1
What the Heck is an Object anyway? File Object G:\folder1\Dog.jpg File Size ACL Date LAN Access SMB/CIFS/NFS Object ID + Metadata All File info plus: M Lady Beagle Likes Kids LAN/WAN/Mobile HTTP/REST API, SDK 2
Flat vs. Hierarchical 4980cdb2a610105904bf7d 2af0d97a04ce483b0380ea 3
Object Storage Characteristics Linear Scalability Scales to billions of objects Support for large files Object sizes are in TBs No Locking No lock on write or create operations Web friendly Firewall friendly, http, REST accessibility Geo-scale Geo-replicated and distributed Metadata and extensibility Objects can be extended to multiple policies (Immutability, retention, etc ) 4
An Object platform offers A flat namespace of millions of buckets OBJECT Buckets that scale to billions of objects OBJECT OBJECT Geo distribution, protection, access OBJECT OBJECT User meta data as a first-class entity Multi-tenant access & metering Data access via REST/HTTP OBJECT 5
Current & Emerging Use Cases for Object Storage Primary Uses Cases Emerging Use Cases Archiving Inexpensive & Secure Long-term Retention Tamper-proof & WORM Use Cases Active Archiving (Emerging Use) Analytics Storage for Big Data/Hadoop projects Scalable storage across multiple architectures Cloud Highly Available Inexpensive Infrastructure For Public & Private Cloud (SPs) On-prem Storage With Hybrid Cloud In Mind Tape/NAS augmentati on On-premise tape replacement NAS tiering Content Distributi on Social Media & Collaboration (Incl. Mobile) Distributed Engineering & Scientific Collaboration Insourcing Migrating applications out of public cloud to on premise Source: 451 Research and Others EMC MARKET INTELLIGENCE 6
Object Optimal for Cloud Use Cases Object semantics more aligned with cloud app requirements Object is built for common cloud storage use cases: Archiving (medical images, email, files, audio/video ) Content distribution / publishing Data acquisition and sharing Write-once, read-many (WORM) Simpler object-style semantics more closely align with application s data access and management requirements Applications simply need to manage byte streams + metadata. "File" metaphor is overly complex and poor architectural fit 7
Who are these guys? 8
What do they all have in common? They are all software-based object storage that can be deployed on commodity off-the-shelf components in a scale-out architecture 9
Why Isn t Everything Software? 10
NOTICE ANYTHING? Copyright 2014 2013 EMC Corporation. All rights reserved. 11
NOTICE ANYTHING? Copyright 2014 2013 EMC Corporation. All rights reserved. 12
Why Isn t Everything Software? Because RAS Matters. 13
Why Isn t Everything Software? Because RAS Matters. (Even When It s Really Cheap*) 14
Commodity Hardware Value Proposition Utilize standardized, open technologies and mass market components Individual components provide lower performance, reliability, etc. At sufficient scale, with the right software, the component pool provides superior characteristics 15
Software-defined means Customers can Leverage Commodity Platforms Software-defined Storage Commodity Platforms 16
Deliver Storage On Commodity Platforms ViPR Services BLOCK STORAGE HDFS STORAGE OBJECT STORAGE Commodity Platforms 17
ViPR Services Architecture Cloud-scale storage services on commodity and arrays ViPR Services BLOCK STORAGE HDFS STORAGE OBJECT STORAGE Geo-Replicated Data Protection Active-Active read/write support with strong consistency No single point of failure Performance and efficiency for small and large objects SITE 1 SITE 2 SITE 3 18
Universal Accessibility Support multiple data types on commodity hardware and arrays Hardware-independent services at cloud scale Simultaneous access to underlying data through multiple interfaces, no ETL Comprehensive Data Access S3, Swift, Atmos and Centera CAS Object APIs HDFS compatible with Cloudera, Hortonworks, Pivotal etc. Block via ScaleIO Extensions to APIs Byte-Range updates, Atomic appends, Rich ACLs etc. 19
+ 3 rd party commodity 3 rd Party Commodity + EMC commodity EMC Commodity Hardware 20
ViPR Services ECS Appliance 21
ViPR Services in an Appliance Cloud storage-in-a-box (Object, Block & HDFS) 9%-28% Cheaper Than Public Cloud Built for HyperScale State-of-the-Art Object Storage Self-Service Provisioning Industry s Most Cost- Effective HDFS Platform 22
ECS Appliance Characteristics Use COTS Components Economies of scale Density Optimized Up to 72TB Raw / Rack Unit Saves Power/GB, Real Estate costs, etc. Labor Optimized Manage the cluster, not the devices Maximize Serviceability Protection Efficiency Geo-efficient storage 23
Commodity Innovation ECS Appliance brings hyperscale and RAS to commodity Innovation and commodity are not contradictory! ECS Appliance components are commodity x86 Servers Ethernet Networking SATA Disk Drives Innovation in how they re put together to enable scalability and serviceability! 24
Storage Overhead Optimized data access, protection and efficiency 3.00 2.50 2.67 Overhead 2.00 1.50 1.33 2.00 1.77 1.67 1.60 1.55 1.52 1.00 0.50 0.00 1 2 3 4 5 6 7 8 # of Data Centers 25
ECS Use Cases Modern Application platform Global Content Repository Geo-scale big Data Analytics https://accesspoint.yourcompany.com Turnkey Public, Private Cloud Storage Geo-scale efficiency & access Lower $/GB than public IaaS L.A. Memphis U.K. Broadest application support Complete Cloud Storage Platform 26