Comparing Apache Ignite And Cassandra For HTAP Applications

Comparing Apache Ignite and CassandraFor HTAP ApplicationsDenis MagdaApache Ignite PMC ChairGridGain Product Management 2018 GridGain Systems, Inc.GridGain Company Confidential

Agenda Hybrid Transactional/Analytical Processing Ignite vs. Cassandra– Denormalized Architecture or Collocated Processing?– Row-level Isolation or Distributed Transactions?– Caching or In-Memory Store? Q&A 2018 GridGain Systems, Inc.GridGain Company Confidential

Hybrid Transaction/AnalyticalProcessing 2018 GridGain Systems, Inc.GridGain Company Confidential

“HTAP is an emerging application architecturethat "breaks the wall" between transactionprocessing and analytics. It enables moreinformed and "in business real time" decisionmaking.” (Gartner Inc., 2014) 2018 GridGain Systems, Inc.GridGain Company Confidential

Ignite vs. Cassandra:Denormalized Architecture or Collocated Processing? 2018 GridGain Systems, Inc.GridGain Company Confidential

Cassandra Path: Denormalized Architecture Denormalization Strategy– Pros– Increasing Performance in Favor of Data RedundancyPerformanceCons–Query-Driven Architecture–Easy to Start and Hard to Evolve 2018 GridGain Systems, Inc.GridGain Company Confidential

Let’s Take an Example 2018 GridGain Systems, Inc.GridGain Company Confidential

Start With a Question!Q1: What are the car models produced by a vendor within aparticular time frame (newest first)? 2018 GridGain Systems, Inc.GridGain Company Confidential

Run QueriesQ1: What are the car models produced by a vendor within aparticular time frame (newest first)?Extra Queries Supported Out of the Box 2018 GridGain Systems, Inc.GridGain Company Confidential

Somewhere in Production Q2: What is the number of cars of a specific modelproduced by a vendor? 2018 GridGain Systems, Inc.GridGain Company Confidential

Let’s reuse existing table!Q2: What is the number of cars of a specific modelproduced by a vendor? 2018 GridGain Systems, Inc.GridGain Company Confidential

Let’s reuse existing table!Q2: What is the number of cars of a specific modelproduced by a vendor?InvalidRequest: code 2200 [Invalid query]message "PRIMARY KEY column"car model" cannot be restricted (precedingcolumn "production year" is not restricted)" 2018 GridGain Systems, Inc.GridGain Company Confidential

Q2:Solution: Create New TableQ1: 2018 GridGain Systems, Inc.GridGain Company Confidential

Ignite Path: Affinity Collocation Related Data Collocation–Countries and Cities, Vendors and Cars–On a Single Cluster NodeCollocated Processing–Efficient SQL JOINS–Map-Reduce for Fast Data–Machine Learning w/o ELT–Reduced Network Traffic - Better Performance 2018 GridGain Systems, Inc.GridGain Company Confidential

Collocated Data Distribution 2018 GridGain Systems, Inc.GridGain Company Confidential

Collocated JOINs21321. Initial Query2. Query execution over local data3. Reduce multiple results in one 2018 GridGain Systems, Inc.GridGain Company Confidential

Machine Learning: Genetic AlgorithmsBiological EvolutionSimulationChromosome and Genes ClusterF1, C1, M1DURABLE MEMORYF F1 F2C C1 C2M M1 M2DURABLE MEMORYF2, C2, M2F Fitness CalculationC CrossoverM Mutation 2018 GridGain Systems, Inc.Collocated ComputationGridGain Company ConfidentialIgnite Cluster

Ignite vs. Cassandra:Row-level Isolation or Distributed Transactions? 2018 GridGain Systems, Inc.GridGain Company Confidential

Cassandra Path: Lightweight Transactions Cassandra is Eventually Consistent DB– With Tunable ConsistencyLightweight Transactions–Aka. Compare and Set Transactions–Row-Level IsolationApplicability–Atomic and linearizable updates of a record–Prevent duplications on inserts 2018 GridGain Systems, Inc.GridGain Company Confidential

Ignite Path: ACID Transactions Distributed ACID Transactions– Pessimistic/Optimistic2 Phase Commit–From RAM to disk Deadlock-free Transactions Applicability–No limitations. Classic Transactions. 2018 GridGain Systems, Inc.GridGain Company Confidential

Ignite vs. Cassandra:Caching or In-Memory Store? 2018 GridGain Systems, Inc.GridGain Company Confidential

Caching Cassandra Off-heap Options–Page Cache–Row Cache–Bloom FiltersJava Heap Options–Key Cache–DSE In-Memory 2018 GridGain Systems, Inc.GridGain Company Confidential

Ignite Memory-Centric StorageAutomaticDefragmentationMemory-Centric StoragePredictable memoryconsumptionDURABLE MEMORYDURABLE MEMORYDURABLE MEMORYFully Transactional(Write-Ahead Log)InstantaneousRestarts 2018 GridGain Systems, Inc.Off-heap Removesnoticeable GC pausesStores Supersetof DataServer NodeServer NodeGridGain Company ConfidentialServer Node

YCSB: Average Load 2018 GridGain Systems, Inc.GridGain Company Confidential

YCSB: Ramping Up Load 2018 GridGain Systems, Inc.GridGain Company Confidential

Caching Cassandra With Ignite No rip-and-replace– Automatic Read/Write-Through– Key-Value OnlyDistributed SQL– Keep CassandraOver Ignite DataACID Transactions–Ignite layer 2018 GridGain Systems, Inc.GridGain Company Confidential

Summary: Ignite or Cassandra? Simplified Architecture– Denormalization vs Affinity CollocationData Consistency and Transactions–Ignite: ACID Transactions–Cassandra: Row-level IsolationIn-Memory Store and Performance–Ignite: read intensive and mixed workloads–Cassandra: write intensive workloads (big load) 2018 GridGain Systems, Inc.GridGain Company Confidential

Any Questions?Thank you for joining us. Follow the gridgain.com#apacheignite#gridgain#dmagda 2018 GridGain Systems, Inc.GridGain Company Confidential

Summary: Ignite or Cassandra? Simplified Architecture - Denormalizationvs Affinity Collocation Data Consistency and Transactions - Ignite: ACID Transactions - Cassandra: Row-level Isolation In-Memory Store and Performance - Ignite: read intensive and mixed workloads - Cassandra: write intensive workloads (big load)