etadatasynchronization ack cyfronet agh kraków, …cgw'18, october 22-24, krakÓw. c....

26
METADATA SYNCHRONIZATION PROTOCOL FOR A DECENTRALIZED NETWORK OF DATA PROVIDERS Łukasz Opioła | Łukasz Dutka | Renata G. Słota | Jacek Kitowski ACK CYFRONET AGH, Krak ów, Poland AGH University of Science and Technology, Krak ów, Poland Faculty of Computer Science, Electronics and Telecommunications Department of Computer Science

Upload: others

Post on 11-Jun-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

METADATA SYNCHRONIZATIONPROTOCOL FOR ADECENTRALIZED NETWORK OFDATA PROVIDERS

Łukasz Opioła | Łukasz Dutka | Renata G. Słota | Jacek Kitowski

ACK CYFRONET AGH, Kraków, Poland

AGH University of Science and Technology, Kraków, PolandFaculty of Computer Science, Electronics and TelecommunicationsDepartment of Computer Science

Page 2: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

Global data access for modern science

Challenges of global data access

Metadata synchronization protocol - requirements

Proposed concept of metadata synchronization protocol

Conclusions

1

2

3

4

5

AGENDA

Page 3: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

GLOBAL DATA ACCESS FOR MODERN SCIENCE

GROUP 1

GROUP 2

GROUP 3

Page 4: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Page 5: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Transparent data access

Bob

Bob

Page 6: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Transparent data access

Cross-border collaboration

BobAlice

Bob

Page 7: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Transparent data access

Cross-border collaboration

Security & efficiency BobAlice

Page 8: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Transparent data access

Cross-border collaboration

Security & efficiency

Openness of the network

BobAlice

GROUP 4

James

Page 9: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Transparent data access

Cross-border collaboration

Security & efficiency

Openness of the network

Autonomy (lack of trust)

GROUP 4

! !!

!

!

!

Page 10: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

CHALLENGES OF GLOBAL DATA ACCESS

GROUP 1

GROUP 2

GROUP 3

Transparent data access

Cross-border collaboration

Security & efficiency

Openness of the network

Autonomy (lack of trust)

Reflecting existing hierarchiesGROUP 4

Page 11: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

METADATA SYNCHRONIZATION PROTOCOL

Page 12: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

GROUP 1

GROUP 2

GROUP 3

GROUP 4

METADATA SYNCHRONIZATION

Page 13: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

Decentralized (retaining autonomy)

Globally consistent

Scalable – hundreds of providers

Secure – decentralized AAI*

Location discovery mechanisms

METADATA SYNCHRONIZATION - REQUIREMENTS

* AAI – authentication and authorization infrastructure

Page 14: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION

Metadata synchronization protocolbased on

hybrid, multi-tier architecture

Page 15: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

Gathers knowledge using metadata sync

Manages data on underlying storages

PROPOSED SOLUTION – PROVIDER SCOPEProvider service – basic building block

AAI – Authentication and Authorization Infrastructure

Page 16: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – ZONE SCOPEZone service – oversees a group of providers

Gathers and serves metadata

Trusted authority & mediator

AuthN & AuthZ center

Macaroon based tokens

Reflects existing hierarchies

Page 17: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – GLOBAL SCOPECooperation between Zones – hybrid architecture

Local authority center

Decentralized, P2P backbone

ZONE A ZONE B

PROVIDER 1 PROVIDER 2 PROVIDER 3

Local authority center

Page 18: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – METADATA SYNCMetadata handled by the Zone of origin – local AuthN & AuthZ center

CREATE / UPDATE / DELETE

READ

AAI AAI

CREATE / UPDATE / DELETE

CREATE / UPDATE / DELETE

READ

READ

ZONE A ZONE B

No concurrent modification conflicts Metadata sync overheads evenly distributed

PROVIDER 1 PROVIDER 2 PROVIDER 3

Page 19: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – METADATA SYNCMulti-tier client-server architecture

PROVIDER 1

PROVIDER 2

PROVIDER 3

ZONE A ZONE B

web interface web interfacenative client

Page 20: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – METADATA SYNCPublish / subscribe and multi-tier caching

FETCH & SUBSCRIBE

USER META

CACHE

CACHE CACHE

CACHE

CACHEZONE A ZONE B

PROVIDER 1

PROVIDER 2

PROVIDER 3

web interface web interfacenative client

FETCH & SUBSCRIBE

FETCH & SUBSCRIBE

FETCH & SUBSCRIBE

FETCH & SUBSCRIBEUSER META

USER META

USER META

USER META

Vast performance improvement

Page 21: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – METADATA SYNCPublish / subscribe and multi-tier caching

PUBLISHPUBLISH

PUBLISH

PUBLISH

PUBLISH

CACHE

CACHE CACHE

CACHE

CACHEZONE A ZONE B

PROVIDER 1

PROVIDER 2

PROVIDER 3

web interface web interfacenative client

MODIFIED

CACHE

CACHE CACHE

CACHE

CACHE

Eventual consistency

Page 22: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

PROPOSED SOLUTION – AAIMacaroon-based decentralized authorization

TOKEN

location https://zone-a.com

public bd3f798c749b

<caveats> …

sig a7e108fa16baeffab0

ZONE A ZONE B

DISCOVERYMECHANISM

PROVIDER 1

PROVIDER 2

PROVIDER 3

web interface web interfacenative client

FETCH & SUBSCRIBE

FETCH & SUBSCRIBE

FETCH & SUBSCRIBE

Page 23: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

A) Open

B) Restricted

C) Isolated

PROPOSED SOLUTION – ZONE MODE

Page 24: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

METADATA SYNC PROTOCOL IN GLOBAL DATA ACCESS

ZONE 1

ZONE 2

ZONE 3

GROUP 1

GROUP 2

GROUP 3

GROUP 4ZONE 4

Page 25: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

Zone service acting as central authority and SP serverSynchronization Protocol (SP) for single zone scopeMacaroon based AAI for single zone scopeData provider service employing the SPCross Zone cooperation supportSync protocol supporting global, cross zone scale

CONCLUSIONS Global data access can be achieved by creating a decentralized network of data providers

We propose a metadata synchronization protocol for such network, based on hybdrid, multi-tier architecture with P2P backbone

Proposed concept is being implemented in Onedata, a distributed virtual file system

Page 26: ETADATASYNCHRONIZATION ACK CYFRONET AGH Kraków, …CGW'18, October 22-24, KRAKÓW. C. HALLENGES OF GLOBAL DATAACCESS. GROUP 1. GROUP 2. GROUP 3 Transparent data access Cross-border

CGW'18, October 22-24, KRAKÓW

THANK YOU

https://onedata.org

The authors gratefully acknowledge the financial support of this work from AGH-UST