Ground up Introduction to In-Memory Data (Grids)

Size: px

Start display at page:

Download "Ground up Introduction to In-Memory Data (Grids)"

Jason Paul
8 years ago
Views:

1 Ground up Introduction to In-Memory Data (Grids) QCON 2015 NEW YORK, NY 2014 Hazelcast Inc.

2 Why you here? 2014 Hazelcast Inc.

3 Java Developer on a quest for scalability frameworks Architect on low-latency projects NoSQL practitioner IMDG newbie Data enthusiast

4 Who am I? 2014 Hazelcast Inc.

6 About me Viktor Solutions Architect at Hazelcast International Java Conferences speaker Co-author of O Reilly s «Enterprise Web Development» Co-organizer of User Groups Princeton JUG Hazelcast User Group NYC

com/gamussa/ Co-author of O Reilly s «Enterprise Web Development» Co-organizer of User

7 Agenda 2014 Hazelcast Inc.

8 The In-Memory landscape you need to know Distributed Data Applications Making In-memory Reliable, Scalable and Durable QA

9 Why In-Memory Data? 2014 Hazelcast Inc.

10 Problem 2014 Hazelcast Inc.

11 Anatomy of a JAVA Application Operating System JVM RAM Application JVM Heap Disk

12 Anatomy of an Application: Tiered Data Storage Tiered Storage Speed (TPS) 2,000,000+ JVM In-Memory Heap Store Disk Size (GB) 2 1,000, ,000 Local Off-Heap Store *** Local Disk Store (Re-startable) 1,000 10,000+ 1,000s External Data Source (e.g., Database)

1,000,000 100,000 Local Off-Heap Store *** Local Disk Store

13 Common Enterprise Application Landscape Load Balancer Load Balancer Load Balancer Load Balancer App Server App Server Application Server Application Application JVM Heap JVM Heap JVM Heap App Server App Server Application Server Application Application JVM Heap JVM Heap JVM Heap App Server App Server Application Server Application Application JVM Heap JVM Heap JVM Heap App Server App Server Application Server Application Application JVM Heap JVM Heap JVM Heap J D B C J D B C J D B C J D B C J D B C J D B C Network J D B C J D B C J D B C J D B C J D B C J D B C Database Databases Database Databases Database Databases Database Databases

Application Application JVM Heap JVM Heap JVM Heap App Server App Server Application Server Application Application JVM Heap JVM Heap JVM Heap J D B C J D B C J

14 Performance & Scalability Challenges JVM memory constraints Heap size limitations GC pauses and tuning Handling peak loads Swapping CPU limitations Limited processor power Many applications sharing CPU cycles CPU intensive processes

Handling peak loads Swapping CPU limitations Limited

15 Performance & Scalability Challenges Reliance on the database All transactions go through the database Expensive and complex to scale Reliance on DBA(s) to maintain and tune Network I/O Bandwidth limitations Latency State-aware cluster Session replication

to scale Reliance on DBA(s) to maintain and tune Network I/O

16 Even More Common Enterprise Application Landscape Load Balancer Load Balancer Load Balancer Load Balancer Enterprise Java Dotnet Devices Phones Tablets Web HTML5 Javascript Rich Clients Next Big Trend???Universal Data Transfer Solution?? Database Databases Database Databases Database Databases Database Databases

HTML5 Javascript Rich Clients Next Big Trend???Universal Data Transfer Solution?

17 In-Memory Data Lanscape 2014 Hazelcast Inc.

18 Messaging Middleware NoSQL DATABASES SQL DATABASES IMDG In-Memory Computing

19 Why In-Memory Now? 70 Network 8 RAM per Commodity Server μs 4 2 TB 0 Fast Ethernet DDR Infiniband 100 GbE Addressable Memory 32 4GB 64 18EB

20 Distributed Data Applications 2014 Hazelcast Inc.

21 Cache-as-a-Service Application Scaling Database Caching Distributed Computing Reactive / Smart Clients

22 Cache-as-a-Service

23 Durability of application data

24 Application Scaling Elastic Scalability (just add more servers and they cluster automatically) Super Speeds (in-memory transaction speeds) High Availability (can deploy in backup pairs or even WAN replicated) Fault Tolerance (no single-point-of-failure) Cloud Readiness (deploy right into EC2, GCE, Docker)

25 Database Caching

26 Distributed Computing

27 Reactive Clients The clients (part of an application) enable you to do the operations without being a member of the grid. Clients can be: smart: this means that they immediately can send an operation like map.get(key) to the member that owns that specific key. dumb: it will connect to a random member in the cluster and send requests to this member. This member then needs to send the request to the correct member. High availability When the connected cluster member dies, client will automatically switch to another live member.

28 Making In-Memory Reliable, Scalable and Durable 2014 Hazelcast Inc.

29 Data Distribution

30 Replication vs. Partitioning Replication - Copying an entire dataset onto multiple servers. Used for improving speed of access to reference records such as master data. Partitioning - Splitting up a large monolithic dataset into multiple smaller sets based on data cohesion.

32 Data Partitioning

33 Data Partitioning

34 Data Partitioning

35 Rebalance Data On New Node 2014 Hazelcast Inc.

36 Distributed Maps Fixed number of partitions (default 271) Each key falls into a partition partitionid = hash(keydata)%partition_count Partition ownerships are reassigned upon membership A B C

37 New Node Added A B C D

38 Migration A B C D

39 Migration A B C D

40 Migration A B C D

41 Migration A B C D

42 Migration A B C D

43 Migration A B C D

44 Crash Migration Complete A B C D

45 Data Safety When Node Dies 2014 Hazelcast Inc.

46 Node Crashes A B C D Crash

47 Backups are Restored A B C D Crash

48 Backups are Restored A B C D Crash

49 Backups are Restored A B C D Crash

50 Backups are Restored A B C D Crash

51 Data is Recovered from backup A B C D Crash

52 Data is Recovered from backup A B C D Crash

53 Backup for Recovered Data A B C D Crash

54 Backup for Recovered Data A B C D Crash

55 Backup for Recovered Data A B C D Crash

56 All Safe A C D

57 Scalability: Up or Out? 2014 Hazelcast Inc.

59 Scale Out

60 Durability Persist to or read from any external data store (SQL database, NoSQL data stores, REST service, EIS, file, etc.) Durability of application data (user sessions, shopping carts, master data, configs) Write through and write behind

61 High Density Caching

62 HD vs Heap Heap: 0 MB Native, 4000 MB Heap, 9 Major GC (4900 ms), 31 Minor GC (4217 ms) HD Memory v1:3000 MB Native, 600 MB Heap, 3 Major GC (2800 ms), 100 Minor GC (2018 ms) HD Memory v2:3300 MB Native, 57 MB Heap, 0 Major GC (0 ms), 356 Minor GC (349 ms)

63 Do I really need NoSQL? 2014 Hazelcast Inc.

64 NoSQL Landscape

65 QA 2014 Hazelcast Inc.

66 Bonus! Go to enter hazelcastic and get your t-shirt!

67 Question, please?

In Memory Accelerator for MongoDB

In Memory Accelerator for MongoDB Yakov Zhdanov, Director R&D GridGain Systems GridGain: In Memory Computing Leader 5 years in production 100s of customers & users Starts every 10 secs worldwide Over 15,000,000