OceanStore : An Architecture for Global-Scale Persistent Storage Jaewoo Kim, Youngho Yi, Minsik Cho.

Slides:



Advertisements
Similar presentations
What is OceanStore? - 10^10 users with files each - Goals: Durability, Availability, Enc. & Auth, High performance - Worldwide infrastructure to.
Advertisements

SPORC: Group Collaboration using Untrusted Cloud Resources Ariel J. Feldman, William P. Zeller, Michael J. Freedman, Edward W. Felten Published in OSDI’2010.
Storage management and caching in PAST Antony Rowstron and Peter Druschel Presented to cs294-4 by Owen Cooper.
Precept 6 Hashing & Partitioning 1 Peng Sun. Server Load Balancing Balance load across servers Normal techniques: Round-robin? 2.
Pond: the OceanStore Prototype CS 6464 Cornell University Presented by Yeounoh Chung.
1 Accessing nearby copies of replicated objects Greg Plaxton, Rajmohan Rajaraman, Andrea Richa SPAA 1997.
OceanStore: An Infrastructure for Global-Scale Persistent Storage John Kubiatowicz, David Bindel, Yan Chen, Steven Czerwinski, Patrick Eaton, Dennis Geels,
Option 2: The Oceanic Data Utility: Global-Scale Persistent Storage John Kubiatowicz.
OceanStore: An Architecture for Global-Scale Persistent Storage John Kubiatowicz University of California at Berkeley.
Scalable and Distributed Similarity Search in Metric Spaces Michal Batko Claudio Gennaro Pavel Zezula.
OceanStore An Architecture for Global-scale Persistent Storage By John Kubiatowicz, David Bindel, Yan Chen, Steven Czerwinski, Patrick Eaton, Dennis Geels,
An Authentication Service Against Dishonest Users in Mobile Ad Hoc Networks Edith Ngai, Michael R. Lyu, and Roland T. Chin IEEE Aerospace Conference, Big.
EEC-681/781 Distributed Computing Systems Lecture 3 Wenbing Zhao Department of Electrical and Computer Engineering Cleveland State University
Large Scale Sharing GFS and PAST Mahesh Balakrishnan.
The Oceanic Data Utility: (OceanStore) Global-Scale Persistent Storage John Kubiatowicz.
Gnutella, Freenet and Peer to Peer Networks By Norman Eng Steven Hnatko George Papadopoulos.
presented by Hasan SÖZER1 Scalable P2P Search Daniel A. Menascé George Mason University.
OceanStore: Data Security in an Insecure world John Kubiatowicz.
OceanStore Theoretical Issues and Open Problems John Kubiatowicz University of California at Berkeley.
Freenet A Distributed Anonymous Information Storage and Retrieval System I Clarke O Sandberg I Clarke O Sandberg B WileyT W Hong.
OceanStore: An Architecture for Global-Scale Persistent Storage Professor John Kubiatowicz, University of California at Berkeley
Concurrency Control & Caching Consistency Issues and Survey Dingshan He November 18, 2002.
Or, Providing Scalable, Decentralized Location and Routing Network Services Tapestry: Fault-tolerant Wide-area Application Infrastructure Motivation and.
OceanStore An Architecture for Global-Scale Persistent Storage Motivation Feature Application Specific Components - Secure Naming - Update - Access Control-
University of Pennsylvania 11/21/00CSE 3801 Distributed File Systems CSE 380 Lecture Note 14 Insup Lee.
OceanStore: An Architecture for Global - Scale Persistent Storage John Kubiatowicz, David Bindel, Yan Chen, Steven Czerwinski, Patric Eaton, Dennis Geels,
Metadata Issues in a Cryptographic File System David Bindel IRAM/ISTORE/OceanStore Retreat.
Presented by: Alvaro Llanos E.  Motivation and Overview  Frangipani Architecture overview  Similar DFS  PETAL: Distributed virtual disks ◦ Overview.
Tapestry GTK Devaroy (07CS1012) Kintali Bala Kishan (07CS1024) G Rahul (07CS3009)
Hands-On Microsoft Windows Server 2008 Chapter 5 Configuring, Managing, and Troubleshooting Resource Access.
Arnold N. Pears, CoRE Group Uppsala University 3 rd Swedish Networking Workshop Marholmen, September Why Tapestry is not Pastry Presenter.
Cooperative File System. So far we had… - Consistency BUT… - Availability - Partition tolerance ?
Jan 17, 2001CSCI {4,6}900: Ubiquitous Computing1 Announcements I will be out of town Monday and Tuesday to present at Multimedia Computing and Networking.
Content Overlays (Nick Feamster). 2 Content Overlays Distributed content storage and retrieval Two primary approaches: –Structured overlay –Unstructured.
Failure Resilience in the Peer-to-Peer-System OceanStore Speaker: Corinna Richter.
Pond: the OceanStore Prototype Sean Rhea, Patric Eaton, Dennis Gells, Hakim Weatherspoon, Ben Zhao, and John Kubiatowicz University of California, Berkeley.
OceanStore: An Infrastructure for Global-Scale Persistent Storage John Kubiatowicz, David Bindel, Yan Chen, Steven Czerwinski, Patrick Eaton, Dennis Geels,
© Oxford University Press 2011 DISTRIBUTED COMPUTING Sunita Mahajan Sunita Mahajan, Principal, Institute of Computer Science, MET League of Colleges, Mumbai.
Distributed Architectures. Introduction r Computing everywhere: m Desktop, Laptop, Palmtop m Cars, Cellphones m Shoes? Clothing? Walls? r Connectivity.
Storage Management and Caching in PAST A Large-scale persistent peer-to-peer storage utility Presented by Albert Tannous CSE 598D: Storage Systems – Dr.
1 More on Plaxton routing There are n nodes, and log B n digits in the id, where B = 2 b The neighbor table of each node consists of - primary neighbors.
OceanStore: An Architecture for Global- Scale Persistent Storage.
Freenet “…an adaptive peer-to-peer network application that permits the publication, replication, and retrieval of data while protecting the anonymity.
Computer Networking P2P. Why P2P? Scaling: system scales with number of clients, by definition Eliminate centralization: Eliminate single point.
Plethora: Infrastructure and System Design. Introduction Peer-to-Peer (P2P) networks: –Self-organizing distributed systems –Nodes receive and provide.
POND: THE OCEANSTORE PROTOTYPE S. Rea, P. Eaton, D. Geels, H. Weatherspoon, J. Kubiatowicz U. C. Berkeley.
Peer to Peer Network Design Discovery and Routing algorithms
Large Scale Sharing Marco F. Duarte COMP 520: Distributed Systems September 19, 2004.
GPFS: A Shared-Disk File System for Large Computing Clusters Frank Schmuck & Roger Haskin IBM Almaden Research Center.
1 Plaxton Routing. 2 History Greg Plaxton, Rajmohan Rajaraman, Andrea Richa. Accessing nearby copies of replicated objects, SPAA 1997 Used in several.
W4118 Operating Systems Instructor: Junfeng Yang.
CS791Aravind Elango Maintenance-Free Global Data Storage Sean Rhea, Chris Wells, Patrick Eaten, Dennis Geels, Ben Zhao, Hakim Weatherspoon and John Kubiatowicz.
Hands-On Microsoft Windows Server 2008 Chapter 5 Configuring, Managing, and Troubleshooting Resource Access.
Data Management on Opportunistic Grids
CS 268: Lecture 22 (Peer-to-Peer Networks)
Option 2: The Oceanic Data Utility: Global-Scale Persistent Storage
OceanStore: An Architecture for Global-Scale Persistent Storage
CHAPTER 3 Architectures for Distributed Systems
Plethora: Infrastructure and System Design
Accessing nearby copies of replicated objects
Distributed P2P File System
OceanStore: Data Security in an Insecure world
A Redundant Global Storage Architecture
OceanStore: An Architecture for Global-Scale Persistent Storage
CSE 542: Operating Systems
CSE 542: Operating Systems
Mid term grades Mean = 48.59, Median = 48.5, Min = 40, Max = 56.
Review Stateless (NFS) vs Statefull (AFS)
Content Distribution Network
Outline for today Oceanstore: An architecture for Global-Scale Persistent Storage – University of California, Berkeley. ASPLOS 2000 Feasibility of a Serverless.
Presentation transcript:

OceanStore : An Architecture for Global-Scale Persistent Storage Jaewoo Kim, Youngho Yi, Minsik Cho

What is P2P storage Distributed File System (DFS) in a P2P manner Allows users to share storages or files – Combined storage capacity of a network of storage devices (peers) Pros – Scalable storage space, resource, persistent storage Cons – Trust, Consistency, Security P2P storage

Introduction OceanStore is A utility infrastructure which consumers pay a monthly fee Span the globe and highly-available from anywhere in the network Continuous access to persistent information Untrusted servers

Two Unique Goals The Ability to be constructed from an untrusted infrastructure Server crashes, leaks information Data is protected by redundancy and cryptographic Responsible party Support of nomadic data Locality is important in a large system Promiscuous caching: distinction between OceanStore and NFS Introspective monitoring: discover tactic relationship – Prevent denial of service attack, regional outages

System Overview Fundamental Unit: Object (e.g. directories, files) Persistent Object: named by GUID (Globally Unique, Identifier) Replicated and stored on multiple servers Provide availability of network partitions, and durability against failure, attack Replica is located by Probabilistic algorithm Deterministic algorithm (if probabilistic algorithm fails) Modified through update Every update create a new version Exist in two forms Active(Latest Version) Archival(Permanent, read-only)

Naming & Access Control Objects are identified by GUID – GUID Thought of as a pseudo-random, fixed-length bit string Secure hash of the owner’s key and some human-readable name Two primitive types of access control – Readers Encrypt all data in the system -> distribute encryption key to users with read permission – Writers Require that all writes be signed -> Server/Client can verify them against an ACL (Access Control List)

Data Location and Routing Two phases of routing 1.Probabilistic algorithm Fully distributed and uses a constant amount of storage per server Attenuated Bloom filter is used Routes to entities rapidly if the entities are in the local vicinity 2.Global deterministic algorithm Randomized data structure with locality properties Used when probabilistic algorithm fails

Probabilistic Algorithm An attenuated Bloom filter is used for probabilistic algorithm An attenuated Bloom filter of depth D can be viewed as an array of D normal Bloom filters – The first Bloom filter is a record of the objects contained locally on the current node – The ith Bloom filter is the union of all Bloom filters for all of the nodes a distance i through any path from the current node

OceanStore: Probabilistic Query Example N N N N N N N N OR 11011

N N N N N N N N Query: N1 local bloom filter: Query: Query: N1 neighbor filter: (Depth 1) Query: N1 neighbor filter: (Depth 2) Query: N2 local bloom filter: Query: N2 neighbor filter: Query: OceanStore: Probabilistic Query Example

Global deterministic algorithm The global algorithm is used when the probabilistic algorithm fails. A variation on Plaxton randomized hierarchical distributed data structure – Every server in the system is assigned a random node-ID. – These node-IDs are then used to construct a mesh of neighboring links. – Each link is labeled with a level number that denotes the stage of routing that uses this link.

Global deterministic algorithm - Example

Update Model Strategy: Conflict resolution, adopted from Bayou system – Dividing by blocks – Updating the modified block – Commiting the modified block (Example: conflict resolution in SVN, Git) Synchronization Problem with Conflict resolution – Untrusted Infrastructure Which server is responsible to update the modified data? – Encrypted Data How conflict resolution is implemented with encrypted data?

File Operations Accessing - Position-Based Cipher – Comparing Version – Comparing Size – Comparing Blocks – Searching Manipulating - Pointers and Encrypted Data Blocks – Replacing Blocks – Inserting Blocks – Deleting Blocks – Appending Data

Serializing Updates

Introspection Cluster Recognition – It detects clusters of strongly-related objects Replica Management – The parent of a replica tracks locally available resources – It can create additional floating replicas to alleviate load

Thank you