Blockchains and databases a new era in distributed computing

Links to Videos, Slides, Bibliography & Twitter Handles @ http://bit.ly/CMbcDB
Blockchains and Databases:
A New Era in Distributed Computing
C. Mohan
IBM Fellow
IBM Almaden Research Center, San Jose, USA
Distinguished Visiting Professor
Tsinghua University, Beijing, China
http://bit.ly/CMwIkP
Twitter, WeChat: seemohan
InfinIT, Aalborg University
Aalborg, Denmark, 25 April 2018

@seemohan2C. Mohan, InfinIT Aalborg, 2018-04-25
Agenda
Goal: Introduce practically relevant private/permissioned blockchains (BCs) to
non-techies AND techies; Also, get into details of many private blockchain systems.
• Origin of Blockchains (BCs)
• Related Distributed Systems/Databases Topics
• Evolution: Smart Contracts, Private BCs, …
• Consortia Approach to Development of Systems
• Applications: Production, PoCs, …
• Market Scene
• Benchmarks
• Architectural Choices and Relationship to DB Replication
• Technical Details of Representative Systems:
Enterprise Ethereum, Hyperledger Fabric & Composer, R3 Corda, Coco Framework,
BigchainDB, Sawtooth, Ripple
• Futuristic Topics

Blockchain (BC)
• Origin in digital currencies, in particular Bitcoin (Satoshi Nakamoto, 2008) – anonymity, open/public/permissionless environment
– awful performance (7 TPS, 10 minutes response time)
• Numerous organizations across the world working on various aspects of it:
security, consensus, database, benchmarks, verification, …
• Banks, regulators, universities, startups, big technology companies, services companies, governments, … individually or as part of
consortia
• February 2017: First production deployment of BC technology by IBM & Northern Trust in Guernsey for administration of private equity fund
managed by Unigestion – Hyperledger Fabric 0.6
• 4/2017: China’s Tencent announces TrustSQL
• July 2017: Hyperledger Fabric 1.0 Released
• Hyperledger Fabric on IBM Cloud - IBM Blockchain Platform (formerly HSBN) on highly secure Linux on mainframes (System Z) with security
hardware – announced August 2017 – available in Dallas, London, Frankfurt, Sao Paolo, Tokyo, Toronto
• 8/2017: Microsoft published Coco Framework
• 10/2017: Oracle announced Blockchain Cloud Service (BCS) - Fabric 1.0 based
• 10/2017: China’s Baidu joined Hyperledger as a Premium Member & 1/2018: Announces BaaS offering
• 3/2018: Hyperledger Caliper Benchmarking Project initiated
• 4/2018: China’s Huawei announces Blockchain Cloud Service for China
• Grand View Research: Global BC Tech Market $7.74B by 2024
• My focus: Private/Permissioned BC Systems!

Bitcoin Blockchain

Bitcoin’s Academic Pedigree
Arvind Narayanan, Jeremy Clark
ACM Queue, August 2017

Distributed Systems
• Distributed operating systems
• Distributed virtual memory
• Message passing in distributed computations and distributed checkpoints
• Clock synchronization and event ordering (e.g., Lamport clocks)
• Byzantine agreement and distributed consensus
• Two phase commit optimizations (e.g., Presumed Abort)
• (Transactional) RPCs and distributed file/object systems
• Asynchronous computation via message queues and pub-sub
• Distributed event-based systems
• Client-server, mobile computing and caching, WWW
• Workflow or business process management systems
• Service Oriented Architecture (SOA)
• Public cloud and hybrid cloud
• …

Data Systems
• Relational DBMSs (e.g., System R) and SQL
• Data consistency, degrees of isolation and fault tolerance
• Distributed databases (e.g., R*) and distributed transactions/queries
• Synchronous and asynchronous replication with primary copy
• Update anywhere (multi-master) replication and eventual consistency
• Stored procedures, user-defined types/functions, data provenance, …
• Data warehousing and parallel DBMSs – OLTP vs OLAP
• Shared Nothing Vs Shared Disks
• Object-oriented databases, XML, schema chaos, data integration, …
• Web2.0-inspired NoSQL, sharding & massive scaling (e.g., Spanner), JSON, …
• Big Data: Map-Reduce, Hadoop, Spark, …
• Data privacy, multitenancy and trans-border data flow restrictions
• Multi data centers and disaster recovery
• …

Port’s
records
Bank’s
records
Airline’s
records
Ocean
Carrier’s
records
Authority's
records
Airport’s
records
… Inefficient, expensive, vulnerable, lack of transparency
Recording of events is becoming much more complex…
Problem Being Solved

Party A’s
records
Bank’s
records
Auditor’s
records
Party B’s
records
Digitally signed,
encrypted
transactions &
ledger
All parties have
same replica
of the ledger
Party A
Bank
Auditor
Party B
Traditional Way Blockchain Way
Clearing
House
… Inefficient, expensive, vulnerable … Consensus, provenance, immutability, finality
Basic Change to Business Processes

http://ibm.biz/Platform_Demo
IBM Blockchain Platform is a fully integrated enterprise-ready
blockchain platform designed to accelerate the development,
governance, and operation of a multi-institution business network
– Developer tools that make use of Hyperledger Composer to
quickly build your blockchain application
– Hyperledger Fabric provides the ledger; managed through a set of
intuitive operational tools
– Governance tools for democratic management of the business
network
– Flexible deployment options, including a highly secure and
performant IBM Cloud environment
IBM Blockchain Platform
Hyperledger Composer
DeveloperTools
IBM Cloud
OperationalTools
GovernanceTools
Blockchain application
Hyperledger Fabric

Actors in a Blockchain Solution

Composer: Workflow of Building a Model

Building Communities in Blockchain Networks
Member Member
Community Based Network
Member Member
Member Member
Founder Directed Network
Member Member
Founder
Member
Member
Consortium Based Network
Member
Founder
Founder
Founder
Examples: Examples:
Founders are equal among other
participants, may include a joint legal entity
among the founders (e.g. – JV)
Examples:
Individual founder in a
position to provide strong
direction
Driven by industry standards
bodies or existing non-
blockchain network owners
IBM Food
Trust

@seemohanC. Mohan, InfinIT Aalborg, 2018-04-25 19
19
19
• Developers create application and smart
contracts (chaincodes)
– Chaincodes are deployed on the network and
control the state of the ledger
– Application handles user interface and
submits transactions to the network which call
chaincodes
• Network emits events on block of
transactions allowing applications to
integrate with other systemscached state,
history, tx
Blockchain
block
txn txn txn txn …
Blockchain
developer
Smart
Contract
Invokes/queriesdevelops
develops
Application
Accesses
event
emits
emits
D
Ledger
‘get’ ‘put, ‘delete’
SDK
Peer
Overview of Application Flow
Database

Blockchain: Distributed Ledger Technology Everest Group

Blockchain Applications
• Track provenance, ownership, relationships & lineage of assets
• Supply Chain – Food Safety (Walmart), Logistics (Maersk)
• Health Data Exchange (FDA)
• Know Your Customer
• Derivatives Processing
• Trade/Channel Finance (IGF)
• Trade Information Warehouse (DTCC)
• Post-Trade Reconciliation/Settlement
• Private Equity Fund Management (Unigestion)
• Syndicated Loans
• Diamond/Valuables Tracking and Protection – Provenance Management (Everledger)
• Cross-Border Payment, Payments for/by Unbanked Populations
• Low volume stock trading (JPX)
“Blockchain Revolution: How the Technology Behind Bitcoin Is Changing Money, Business, and the World”
Don Tapscott, Alex Tapscott, ISBN 978-1101980132.

What?
• Provide trusted source of information and traceability to
improve transparency and efficiency across food network
How?
• Shared ledger for storing digital compliance documentation,
test results and audit certificates network
Benefits
• Reduce impact of food recalls through instant access to
end-to-end traceability data to verify history in food network
& supply chain
• Help address 1 in 10 people sickened and 400K fatalities
worldwide which occur every year from food-born illnesses
Food Trust/Safety

What?
• An open, extensible platform for sharing shipping
events, messages, and documents across all the
actors and systems in the supply chain ecosystem.
How?
• Providing Shared Visibility and Shared State for
Container Shipments
Benefits
• Increase speed and transparency for cross border
transactions through real time access to container
events.
• Reduced cost and increased efficiency through
paperless trade
Provider of interface:
value-add partners
Event publishers
& subscribers
GTD
Platform
Customs
Dashboard
Logistic actors
internal systems
Supply Chain
Visibility systems
Port community
systems
Transportation
management
Shippers
Authorities
Trade Associations
Carriers
Supply Chain
Actors
Terminals
Supply Chain
Management
Global Trade Digitization (GTD)

Dubai Blockchain Strategy
Source: Saeed Al Dhaheri, Etisalat Academy 3/2017

What?
• Create and promote secure, efficient and scalable
exchange of health data using blockchains
How?
• Create an electronic ledger of where and how data is
transferred and exchanged
• Initial trial focus on oncology data
Benefits
• Creating an audit trail through the ledger, healthcare
professionals will be able to:
– hold information leakers accountable
– maintain transparency in what data is going where
– secure weak spots in sharing process
IBM-FDA Partnership

Blockchain Companies/Consortia & Banks
Source: American Banker 2-2017

Global Bitcoin/Blockchain Companies 3/2017
Source: https://www.cbinsights.com/blog/bitcoin-blockchain-startup-global-map/

ICOs and Equity Deals
CB Insights: 2017 saw over 5x more capital deployed in ICOs than in equity financings to
blockchain startups. In total, ICOs raised over $5B across nearly 800 deals in 2017. To put
those numbers into context, equity investors deployed $1B in 215 deals to the sector.

BC Software Stack
Source: Anh Dinh, et al., SIGMOD 2017
Consensus Layer (PBFT, PoW, PoS, POA, etc.)
Smart Contract Execution Engine
(Virtual Machine, Docker, etc.)
Data Model Layer
(LevelDB, RocksDB, etc.)

Blockchain Architecture/Feature Choices
• Cryptocurrencies Vs Generalized Assets
• Permissionless/Public Vs Permissioned/Private
• Byzantine Vs Non-Byzantine fault model
• Consensus approach: PoW, PoA, PoET, PBFT, …
• SQL Vs NoSQL data stores
• Transactional stores Vs Non-transactional stores
• Versioned/Unversioned state database
• On-Chain Vs Off-Chain data
• Parallelism exploitation during different phases of transaction execution
• Pluggable features: consensus protocol, state DB, smart contract language, …
Good Survey Paper: Untangling Blockchain: A Data Processing View of Blockchain Systems, A. Dinh et al.

Database Replication
• Primary log replay at replica – homogeneous systems with full DB replicas,
typically done for disaster recovery (DR) backup
• Log capture generates DML statements from what is logged and apply executes
those statements (e.g., IBM Q Replication)
• Can handle non-determinism and partial replicas
• Requires dependency analysis to leverage parallelism at apply time
• https://www.ibm.com/developerworks/data/roadmaps/qrepl-roadmap.html
• Capture DML statements as issued by application and re-execute them at replica
(e.g., H-Store/VoltDB)
• Cannot handle non-determinism
• Typically, serial execution of transactions
Upfront (fairly random, unoptimized) ordering of transactions in blockchain
systems – leads to all sorts of issues!

Benchmark Framework: BLOCKBENCH (NUS)
• OLTP & OLAP load measured
• Metrics: throughput, latency, scalability, fault
tolerance, security
• Consensus methods: Ethereum (PoW), Fabric
(PBFT), Parity (PoA)
• Old version of Fabric (pre-V1)
• Fabric performs better
• Fabric scales well up to 16 nodes
Source: Anh Dinh, et al., SIGMOD 2017

Hyperledger Caliper
• Allows users to measure performance of a specific blockchain implementation
with a set of predefined use cases
• Will produce reports containing a number of performance indicators, such as
TPS (Transactions Per Second), transaction latency, resource utilization, …
• Intent is for Caliper results to be used by other Hyperledger projects as they
build out their frameworks, and as a reference in supporting the choice of a
blockchain implementation suitable for a user’s specific needs
• Initial contributors: Developers from Huawei, Hyperchain, Oracle, Bitwise,
Soramitsu, IBM and Budapest University of Technology and Economics
• https://www.hyperledger.org/projects/caliper

Ethereum
• Public blockchain system like Bitcoin
• Extends it with Smart Contracts
• Uses PoW for consensus
• Own machine lang & VM
• gas charging!
• Most apps relate to its currency Ether
• Enterprise Ethereum Alliance (EEA): JPMorgan Chase, Microsoft, Intel, Accenture, Banco Santander,
BNY Mellon, ConsenSys, Credit Suisse, ING, Thomson Reuters, UBS, Wipro
• EEA will add confidentiality (Quorum), scalability (pluggable consensus) and
permissioning to Ethereum
• Focus on specification, EntEth 1.0 with Python reference client, benchmarking,
compliance testing and tools
• Develop standards for Ethereum: best practices, security, privacy, scalability,
interoperability
• Quorum from JPMorgan; Support for PBFT added in 7/2017 by AMIS

Quorum Network J.P. Morgan

Hyperledger: A Linux Foundation project
• A collaborative effort created to advance cross-
industry blockchain technologies for business
• Founded February 2016; now more than 230
member organizations
• Open source, open standards, open governance
• Five frameworks and five tools projects
• IBM is a premier member of Hyperledger
www.hyperledger.org

Hyperledger Fabric Project
• Initiated by IBM with IBM open source ledger contribution (Feb 2016)
http://hyperledger-fabric.readthedocs.io/en/latest/
• Significant change in architecture from V0.6 to V1
• Chaincode trust flexibility
• Channel concept for Scalability & Confidentiality enhancement
• Consensus modularity
• Pluggable State DB APIs
• 2 types of peer nodes: Endorsing, non-endorsing/committing
• Used PBFT for consensus before V1
• Other Hyperledger Projects: Iroha, Sawtooth, Composer, Quilt, …
Hyperledger Premier members include: Accenture, Airbus, American Express, Baidu, Change Healthcare, Cisco, CME Group, Deutsche
Borse Group, Daimler, Digital Asset, DTCC, Fujitsu, Hitachi, IBM, Intel, J.P. Morgan, NEC, R3, SAP, Tradeshift and Wanda FFan Technology
Hyperledger Fabric V1 Contributors - Engineers from: Arxan, Cloudsoft, CLS, d20 Technical Services, Depository Trust & Clearing
Corporation (DTCC), Digital Asset, Fujitsu, GE, Gemalto, HACERA, Hitachi, Huawei Technologies, Hyperchain, ImpactChoice, IT People,
Knoldus, Linux Foundation, Netease, Passkit, State Street Bank, SecureKey, IBM, SAP, Thoughtworks and Wanda Group. There were also
contributions from 35 unaffiliated individuals. In total, 159 developers have contributed.

Fabric V1 Architecture
• Elements of the Architecture
• Chaincode (in Go): System and regular ones. Deploy/Invoke latter.
• Assets represented as Key-value pairs in binary and/or JSON form
• State: Versioned KV model - stored in Level DB or CouchDB
• Ledger data: Blockchain has full state, including history
• Transactions
• Kafka for Ordering:
• No Byzantine fault handling
• Done to improve performance
• Pluggable consensus permits other methods
IBM Blockchain Platform and Oracle Blockchain Cloud Service based on it

@seemohanC. Mohan, InfinIT Aalborg, 2018-04-25 54
54
54
Latest written key/values for
use in transaction simulation
Supports keyed queries,
composite key queries, key
range queries
TX
Reads[]
Writes[]
TX
Reads[]
Writes[]
TX
Reads[]
Writes[]
TX
Reads[]
Writes[]
State Database
Transaction Log
Last written key/value
CouchDB (external option)
supports keyed queries,
composite key queries, key
range queries, plus full data
rich queries (beta in 1.0)
blockHash à SegNo + offset
blockNum à SegNo + offset
txId à SegNo + offset
Block index
block
tx array
File System Level DB
Replaceable
tracking history of a key
Key History index
Fabric V1 Ledger

Transaction Execution Overview Fabric V1
• Transaction is sent to the counter-parties
represented by Endorsing Peers on their Channel
• Each Peer simulates transaction execution by
calling specified Chaincode function(s) and signs
result (Read-Write Sets)
• Each Peer may participate in multiple channels
allowing concurrent execution
• Ordering Service accepts endorsed transactions
and orders them according to the plug-in
consensus algorithm then delivers them on the
channel
• All (Committing) peers on channel receive
transactions: on successful validation, commit to
ledger. No chaincode execution.
3 Stage Execution: Endorsement, Ordering, Validation/Commit
Client
• Invoke
Chaincode
• Receive
Read and
Write Sets
• Send tran
to Ordering
Service
• Hear back
final
outcome
Endorsing
Peer
Endorsing
Peer
Committing
Peer
chaincode
chaincode
Ordering
Service
Channel
1
2
3
1 3
3
Channel.SendTransactionProposal (Step 1) and channel.SendTransaction (Step 2)

Architectural Differences
http://bit.ly/HFpapr

DBMS Implications
• Simulation concept requires layer between chaincode and State DB having to
take on analysis of DBMS calls
• Update statements split into two: read part and write part
• Read alone sent to DBMS with modifications to retrieve version #s for items read
• Writes not sent to DBMS but processed and cached locally – doesn’t allow for read
your own write by chaincode transaction
• During Commit phase, read sets validated by retrieving each item’s version #
individually and then, if validation succeeds, writes also done one at a time
• Dealing with phantoms requires reexecution of query during commit phase to be
sure simulation read set same as read set at Commit time
• Chaincode portability across different State DBMSs hard to do
• Lots of open questions and research issues in this area

IBM Blockchain Engagements
Making blockchain real for business with more than 400 engagements and multiple active networks
Trade Finance Pre and Post Trade Complex Risk Coverage
Identity/ Know your customer (KYC) Unlisted Securities/ Private Equity Funds Loyalty Program
Medicated Health Data Exchange Fraud/ Compliance Registry Distributed Energy/ Carbon Credit
Supply Chain Food Safety Provenance/ Traceability

R3 Alliance & Corda
• Barclays, BBVA, Commonwealth Bank of Australia (CBA), Credit Suisse, J.P. Morgan,
State Street, Royal Bank of Scotland, UBS
• Special features for JVM to guarantee deterministic behavior
• Hearn, M. Corda: A distributed ledger, Version 0.5, November 2016.
https://docs.corda.net/_static/corda-technical-whitepaper.pdf
• Nodes backed by RDBMS, ledger data SQL queryable and joinable with private tables
• Corda written in Kotlin (simpler Scala with much better Java interoperability) from
JetBrains – contracts in Kotlin/Java
• Contract execution is deterministic and its acceptance of a transaction is based on the
transaction’s contents alone. A transaction is only valid if the contract of every input
state and every output state considers it to be valid

R3 Corda Vault

Sawtooth (Intel)
• Project of Hyperledger; 1.0 release announced in 1/2018
• Proof of Elapsed Time (PoET) – Consensus Protocol
• Every validator requests a wait time from a trusted function
• Validator with shortest wait time for a particular transaction block is elected leader
• Guaranteed wait time
• Randomness in leader election (~ to lottery algorithm)
• Intended to run in a Trusted Execution Environment (TEE), e.g., Intel’s Software
Guard Extensions (SGX)
• Concept of Transaction Family and Transaction Dependencies
• Transaction Scheduling: Serial or Parallel
• Same block can contain multiple transactions which modify same value!
• https://sawtooth.hyperledger.org/docs/core/releases/latest/contents.html

Microsoft Coco Framework/System
Designed to be open and compatible with any blockchain protocol

Huawei Blockchain Cloud Service
• Announced April 2018
• BaaS offering based on Hyperledger Fabric V1
• An off-chain channel trading system for the Hyperledger Fabric platform, which is highly
efficient and secure
• Achieves a single-channel 2,000+ TPS transaction performance between users
• Initial offering only in China?
• News Articles
https://cryptovest.com/news/huawei-brings-out-blockchain-as-a-service-offering
https://www.coindesk.com/huawei-unveils-hyperledger-based-blockchain-service-platform/
• Whitepaper: https://static.huaweicloud.com/upload/files/pdf/20180411/20180411144924_27164.pdf
Zhu Peiying, Liao Heng, Wu Shuang, William Michael Genovese et al.

Application Flow with RDBMS (In Progress)
• Developers create application
and smart contracts (chaincodes)
– Chaincodes are deployed on the
network and control the state of the
ledger
– Application handles user interface
and submits transactions to the
network which call chaincodes
• Network emits events on block of
transactions allowing applications
to integrate with other systems
Cached State
Blockchain
block
txn txn txn txn …
Blockchain
Developer
Smart
Contract
Invokes/queriesDevelops
Develops
Application
Accesses
Event
Emits
Emits
D
Ledger
‘SELECT’
‘INSERT’, ‘DELETE’, ‘UPDATE’
‘CREATE’, ‘ALTER’, “CALL’
SDK
Peer
RDBMS

Futuristic Topics
• Chaincode portability and power of data APIs
• DBMS enhancements to add BC features
• Standards across BC systems
• Cross channel transactions
• Non-deterministic actions
• Analytics on chaincode data
• Many app design issues
• Design tools for endorsement decisions
• NL contracts -> formal contracts ->
executable contracts
Numerous research possibilities for database and distributed systems
people in this new era of distributed computing!

More Information
Links to Videos, Slides, Bibliography, Twitter Handles
http://bit.ly/CMbcDB
Follow me on
Twitter, WeChat, Instagram: @seemohan
Facebook: http://www.facebook.com/cmohan
LinkedIn: http://www.linkedin.com/in/seemohan/

Blockchains and databases a new era in distributed computing

Recommended

More Related Content

What's hot (20)

Similar to Blockchains and databases a new era in distributed computing (20)

More from InfinIT - Innovationsnetværket for it (20)

Recently uploaded (20)

Blockchains and databases a new era in distributed computing