KEMBAR78
Distributed caching-computing v3.8 | PPTX
Rahul Gupta, profile picture
Uploaded byRahul Gupta
794 views

Distributed caching-computing v3.8

Thinking Distributed: The Hazelcast Way document discusses Hazelcast, an in-memory data grid that provides distributed computing capabilities. It describes how Hazelcast enables scale-out computing, resilience to failures, and an easy programming model. It also outlines Hazelcast's features such as fast performance, persistence, SQL queries, and support for various APIs and languages.

Thinking Distributed: The Hazelcast Way RAHUL GUPTA SR SOLUTIONS ARCHITECT
About Me Senior Solutions Architect Worked with Terracotta In-memory Distributed Systems since 2009 Java Programmer since 1998 rahul@hazelcast.com Rahul Gupta follow me @wildnez
© 2017 Hazelcast Inc. Confidential & Proprietary 3 Hazelcast Reborn
1010110010000101001 1010110010001110101001 10111001001 10101100100001111010100 Challenges of Big Data Management
Why Hazelcast? • Scale-out Computing enables cluster capacity to be increased or decreased on-demand • Resilience with automatic recovery from member failures without losing data while minimizing performance impact on running applications • Programming Model provides a way for developers to easily program a cluster application as if it is a single process • Fast Application Performance enables very large data sets to be held in main memory for real-time performance
© 2017 Hazelcast Inc. Confidential & Proprietary 6 Low-Level Services API Node Engine (Threads, Instances, Eventing, Wait/Notify, Invocation) Partition Management (Members, Lite Members, Master Partition, Replicas, Migrations, Partition Groups, Partition Aware) Cluster Management with Cloud Discovery SPI (AWS, Azure, Consul, etcd, Heroku, IP List, Apache jclouds, Kubernetes, Multicast, Zookeeper) Networking (IPv4, IPv6) On-Heap Store High-Density Memory Store (Intel, Sparc) Serialization (Serializable, Externalizable, DataSerializable, IdentifiedDataSerializable, Portable, Custom) Near Cache HD C++ C#/.NET Python Node.js Scala Memcached Clojure Open Client Network Protocol (Backward & Forward Compatibility, Binary Protocol) REST Java JVM (JDK: 6, 7, 8, Vendors: Oracle JDK, Open JDK, IBM JDK, Azul Zing & Zulu) Hardware/VM/Container (On Premise, AWS, Azure, Docker, Pivotal Cloud Foundry, OpenShift, VMware) Executor Service SQL Query MapReduce AggregationPredicate Entry Processor Map MultiMap Replicated Map Web Sessions (Tomcat/Jetty/Generic) Continuous QueryHD HD Open Source Enterprise Edition Hazelcast IMDG Enterprise HD Edition HD-Enabled Feature WAN (Socket,SolaceSystems:One-way,Multiway,InitnewDataCenter,DRDataCenterRecovery) ManagementCenter (JMX/REST) SecuritySuite (Connection,Encryption,Authentication,Authorization,JAASLoginModule,SocketInterceptor) Engine Clients HD HD HD Queue Topic Ring Buffer Atomic Long Hibernate 2nd Level Cache HD JCache Set Lock/ Sem. HD java.util.concurrent UUID Hot Restart Store (SSD, HDD) HD Storage Operating System (Linux, Oracle Solaris, Windows, AIX, Unix) Operating Env APIs Hyper LogLog RollingUpgrades (RollingClientUpgrades,RollingMemberUpgrades,NoDowntime) HazelcastStriimHotCache (SyncUpdatesfromOracleDB,MSSQLServer,MySQLandNonStopDB) Operations Hazelcast Solution
Ecosystem Traction Dozens of Commercial and Open Source Projects Embed Hazelcast
01001 10101 01010 In Memory Data Computing In Memory Data Messaging++In Memory Data Storage In Memory Data Grid
Business Systems A B C RDBMS Mainframe MongoDB NoSQL REST Scale Hazelcast In-Memory Caching
java.util.concurrent.ConcurrentMap import java.util.concurrent.ConcurrentHashMap; import java.util.concurrent.ConcurrentMap; public static void main(String[] args) { ConcurrentMap<Integer, String> map = new ConcurrentHashMap<>(); map.put(1, "Paris"); map.put(2, "London"); map.put(3, "San Francisco"); String oldValue = map.remove(2); }
Distributed Map import java.util.concurrent.ConcurrentMap; import com.hazelcast.core.Hazelcast; import com.hazelcast.core.HazelcastInstance; public static void main(String[] args) { HazelcastInstance h = Hazelcast.newHazelcastInstance(); ConcurrentMap<Integer, String> map = h.getMap("myMap"); map.put(1, "Paris"); map.put(2, "London"); map.put(3, "San Francisco"); String oldValue = map.remove(2); }
DEMO
Persistence API public class MapStorage implements MapStore<String, User>, MapLoader<String, User> { // Some methods missing ... @Override public User load(String key) { return loadValueDB(key); } @Override public Set<String> loadAllKeys() { return loadKeysDB(); } @Override public void delete(String key) { deleteDB(key); } @Override public void store(String key, User value) { storeToDatabase(key, value); } } <map name="users"> <map-store enabled="true"> <class-name>com.hazelcast.example.MapStorage</class-name> <write-delay-seconds>0</write-delay-seconds> </map-store> </map>
JCache API // Retrieve the CachingProvider which is automatically baced by // the chosen Hazelcast server or client provider CachingProvider cachingProvider = Caching.getCachingProvider(); // Create a CacheManager CacheManager cacheManager = cachingProvider.getCacheManager(); // Cache<String, String> cache = cacheManager // .getCache( name, String.class, String.class ); // Create a simple but typesafe configuration for the cache CompleteConfiguration<String, String> config = new MutableConfiguration<String, String>() .setTypes( String.class, String.class );
JCache API // Create and get the cache Cache<String, String> cache = cacheManager .createCache( "example", config ); // Alternatively to request an already existing cache Cache<String, String> cache = cacheManager .getCache( name, String.class, String.class ); // Put a value into the cache cache.put( "world", "Hello World" ); // Retrieve the value again from the cache String value = cache.get( "world" ); System.out.println( value );
Eviction • Unless deleted, entries remain in the map. • Use eviction policies to prevent OOM situations. • Eviction Triggers run LFU or LRU. time-to-live max-idle-seconds max-size (PER_NODE(number of entries), PER_PARTITION(number of entries), USED_HEAP_SIZE(mb), USED_HEAP_PERCENTAGE(mb)) • Setting eviction-percentage removes that % of entries when eviction is triggered.
Features Description MultiMap Store multiple values against one Key Replicated Map Cluster wide replication, all entries on all nodes. Good for read heavy use cases Near Cache Map Entries on Client/Application local memory RingBuffer Stores data in a ring-like structure, like a circular array with given capacity More Distributed Structures
Java Collection API: Map, List, Set, Queue Jcache High Density Memory Store Hibernate 2nd Level Cache Web Session Replication: Tomcat, Jetty Predicate API: Indexes, SQL Query Persistence: Map/Queue Store & Loader. Write Behind/Through Spring Compliance Transactions: Local & XA WAN & DR Replication IM Data Store (Caching) Features
Java: Will it make the cut? Garbage Collection limits heap usage. G1 and Balanced aim for <100ms at 10GB. Unused Memory 64GB 4GB 4s Heap Java Apps Memory Bound GC Pause Time Available Memory GC Off-Heap Storage No low-level CPU access Java is challenged as an infrastructure language despite its newly popular usage for this Heap
GC Pause 64 Gb JVM 4 Gb JVM 1 4 Gb JVM 2 4 Gb JVM 3 4 Gb JVM 4 4 Gb JVM 5 4 Gb JVM 16 GC Pause 0 GB 64 GB 64 Gb in-memory data time Standard Impediments of Caching
GC Pause 64 Gb JVM 4 Gb JVM 1 4 Gb JVM 2 4 Gb JVM 3 4 Gb JVM 4 4 Gb JVM 5 4 Gb JVM 16 GC Pause 0 Gb 64 Gb 64 Gb in-memory data time 64 Gb HD 2 Gb JVM GC Pause HD Memory Caching with HD Memory
On-Heap Memory Store (Objects Stored as Objects) High-Density Memory Store (Objects Serialized and Stored as Bytes) s.m.Unsafe s.m.Unsafe 2-4GB (Limited by Garbage Collection) 0-1TB (Limited by Machine RAM) Memory Stores •Member •Client (Near Cache) RAM in JVM Process APIs JCache (ICache) Map (IMap) HD Memory
On Heap Vs. High-Density Memory Management On Heap Memory HD Memory 0 MB HD 3.3 GB 3.9 GB Heap Storage 0.6 GB 9 (4900 ms) Major GC 0 (0 ms) 31 (4200 ms) Minor GC 356 (349 ms) Node Used Heap Total Heap Max. Heap Heap Usage Percentage Used Heap: 0.2 GB 192.168.1.10:5701 57 MB 229 MB 910 MB 6.28% Memory Utilization Home Offheap-test Node Used Heap: 3.9 GB 192.168.1.10:5701 3933 MB 4658MB 4653MB 84.45% Memory Utilization Home Used Heap Total Heap Max. Heap Heap Usage Percentage Example: On Heap Memory Example: HD Memory
Hazelcast Servers Hazelcast Server JVM [Memory] A B C Business Logic Data Data Data CE = Compute Engine Result Business / Processing Logic Result TCP / IP Client Client Distributed Computing
IM Distributed Computing Feature HD Cache Dist. Compute Dist. Message Java Concurrency API (Lock, Semaphore, AtomicLong, AtomicReference, Executor Service, Blocking Queue) Entry and Item Listeners Entry Processor Aggregators Map/Reduce Data Affinity Continues Query Map Interceptors Delta Update
Executor Service API public interface com.hazelcast.core.IExecutorService extends java.util.concurrent.ExecutorService HazelcastInstance hz = getHazelcastInstance(); //java.util.concurrent.ExecutorService implementation IExecutorService es = hz.getExecutorService("name"); es.executeOnAllMembers(buildRunnable()); es.executeOnKeyOwner(buildRunnable(), "Peter"); es.execute(buildRunnable()); Map<..> futures = es.submitToAllMembers(buildCallable()); Future<..> future = es.submitToKeyOwner(buildCallable(), "Peter"); es.submitToAllMembers(buildCallable(), buildCallback()); es.submitToKeyOwner(buildCallable(), "Peter", buildCallback());
EntryProcessor API
Lock API HazelcastInstance hz = getHazelcastInstance(); // Distributed Reentrant L​ock l​ock = hz.getLock("myLock"); l​ock.lock(); try { // Do something } finally { l​ock.unlock(); } Distributed Lock
Lock API Pessimistic Locking (IMap)
Lock API Optimistic Locking
Map/Reduce API HazelcastInstance hz = getHazelcastInstance(); Map users = hz.getMap("users"); JobTracker tracker = hz.getJobTracker("default"); KeyValueSource source = KeyValueSource.fromMap(users); Job job = tracker.newJob(source); ICompleteFuture future = job.mapper(new MyMapper()) .reducer(new MyReducer()) .submit(); Map result = future.get();
Aggregations API HazelcastInstance hz = getHazelcastInstance(); Map users = hz.getMap("users"); int sum = users.aggregate( Supplier.all((user) -> user.getSalary()), Aggregations.longSum() );
Hazelcast Distributed Topic Bus Hazelcast Topic Hazelcast Node 1 Hazelcast Node 2 Hazelcast Node 3 MSG Subscribes Delivers Subscribes Delivers Distributed Messaging
Queue Topic (Pub/Sub) Event Listeners Ring Buffers IM Distributed Messaging Features HD Cache Dist. Compute Dist. Message
Queue API interface com.hazelcast.core.IQueue<E> extends java.util.concurrent.BlockingQueue HazelcastInstance hz = getHazelcastInstance(); //java.util.concurrent.BlockingQueue implementation IQueue<Task> queue = hz.getQueue("tasks"); queue.offer(newTask()); queue.offer(newTask(), 500, TimeUnit.MILLISECONDS); Task task = queue.poll(); Task task = queue.poll(100, TimeUnit.MILLISECONDS); Task task = queue.take();
Topic API public class Example implements MessageListener<String> { public void sendMessage { HazelcastInstance hz = getHazelcastInstance(); ITopic<String> topic = hz.getTopic("topic"); topic.addMessageListener(this); topic.publish("Hello World"); } @Override public void onMessage(Message<String> message) { System.out.println("Got message: " + message.getMessageObject()); } }
Data Distribution and Resilience 38
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 change A B C
New Node Added DA B C
Migration DA B C
Migration DA B C
Migration DA B C
Migration DA B C
Migration DA B C
Migration DA B C
Migration Complete DA B C
Data Safety on Node Failure 48
Node Crashes DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Backups Are Restored DA B C
Recovery Is Complete DA C
© 2017 Hazelcast Inc. Confidential & Proprietary 59 Topologies Deployment Architectures
© 2017 Hazelcast Inc. Confidential & Proprietary 60 Embedded Mode Great for microservices, OEM and ops simplification IMDG Application Java API IMDG Application Java API IMDG Application Java API
© 2017 Hazelcast Inc. Confidential & Proprietary 61 Embedded Mode Embedded model, for example in a J2EE container Service and storage in the one JVM Easy for getting started With three nodes, 1/3rd of the data is local, good read performance -> This benefit diminishes as you scale up -> With ten nodes, only 1/10th of the data is local If the application needs to go offline, so does the data
© 2017 Hazelcast Inc. Confidential & Proprietary 62 Client-Server Mode Great for scale-up or scale-out deployments with cluster lifecycle decoupled from app servers Use for non-Java clients Node.js API Application Node.js API Application Node.js API Application Node.js API ApplicationJava API Application .Net API Application C++ API Application C++ API Application C++ API Application C++ API Application .Net API Application .Net API Application .Net API ApplicationJava API Application Java API Application IMDG IMDG IMDG
© 2017 Hazelcast Inc. Confidential & Proprietary 63 Client-Server Mode Follows the traditional Client-Server approach Clients do not change cluster topology Allows you to segregate service from storage Smart clients connect to all cluster members. Operations are routed directly to a member holding the necessary data. -> One network hop Other clients connect to a single cluster member and use it as a proxy to issue requests. -> Two network hops if the proxy doesn’t have the data needed
© 2017 Hazelcast Inc. Confidential & Proprietary 64 Pushing Into New Language Ecosystems Hazelcast.org/Plugins • Fully featured • Hazelcast supported • Open Source • Common Hazelcast Feature Set • Hazelcast supported • Open Source • High Performance • Hazelcast supported • Open Source • Client and Member APIs • Scala 2.11 support • Hazelcast Supported • Open Source • Community supported • Open Source • Fully Featured • Hazelcast supported • Open Source • Industry Leading Performance • High-Density Memory Store Near Cache • Hazelcast Supported • Open Source Feature Matrix: https://hazelcast.org/clients-languages
© 2017 Hazelcast Inc. Confidential & Proprietary 65 Hazelcast IMDG Microservices Deployment Options Microservices deployed embedded with isolated Hazelcast IMDG clusters Maximum service isolation Microservices deployed client server with a shared Hazelcast IMDG cluster Transition to microservices MS 1 IMDG IMDG IMDG MS 1 MS 1 .NET Client MS 1 IMDG IMDG IMDG IMDG .NET Client MS 1 Node.js Client MS 2 Node.js Client MS 2 C++ Client MS 3 Java Client MS 4 Java Client MS 4 MS 2 IMDG IMDG IMDG MS 2 MS 2 MS 3 IMDG IMDG IMDG MS 3 MS 3
© 2017 Hazelcast Inc. Confidential & Proprietary 66 Hazelcast IMDG Microservices Deployment Options Microservices deployed client server with isolated Hazelcast cluster per service Maximum service isolation and non-Java clients .NET Client MS 1 IMDG IMDG IMDG .NET Client MS 1 .NET Client MS 1 .NET Client MS 1 .NET Client MS 1 Python Client MS 2 IMDG IMDG IMDG Python Client MS 2 Python Client MS 2 Python Client MS 2 Python Client MS 2
Same Features – Embedded vs Client-Server // Creating a new Hazelcast node HazelcastInstance hz = Hazelcast.newHazelcastInstance(); // Getting a Map, Queue, Topic, ... Map map = hz.getMap("my-map"); Queue queue = hz.getQueue("my-queue"); ITopic topic = hz.getTopic("my-topic"); //Creating a Hazelcast Client HazelcastInstance client = HazelcastClient.newHazelcastClient(); // Shutting down the node hz.shutdown();
© 2017 Hazelcast Inc. Confidential & Proprietary 68 Hazelcast IMDG – Any Cloud, Anytime Apache jClouds PaaS Environment Integrations Cloud & Cloud Discovery Plugins Deployment Options YARN
© 2017 Hazelcast Inc. Confidential & Proprietary 69 Hazelcast IMDG WAN Replication Deployment Options Independent clusters, sharing selected data to maintain alignment One-way or two-way Choose which data to share, select by type, can de-select records by attributes One-way  keep DR site populated and ready-to-go Two-way  worldwide operations, access your nearest data copy API Applications API Applications API Applications IMDG IMDG IMDG IMDG IMDG IMDG IMDG
70 Hazelcast IMDG Roadmap
© 2017 Hazelcast Inc. Confidential & Proprietary 71 Hazelcast IMDG High Level Roadmap Hi-Density Caching In-Memory Data Grid 2015 2016 JCache | Scalability | HD Memory | Hot Restart | Pre-Load Clients Cloud Discovery SPI | Azure Cloud Foundry | AWS j.u.c. | Low-Latency | Elasticity | Resiliency | Performance | Rolling Upgrades Open Cloud Platform 2017 j.u.s | Jet | Cascading | Streaming | Mesos | Reactive Expanded In-memory Computing 2018
© 2017 Hazelcast Inc. Confidential & Proprietary 72 Recent Releases & Roadmap Release Date Hazelcast IMDG 3.7 August 2016 (Released) Hazelcast IMDG 3.8 February 2017 (Released) Hazelcast IMDG 3.9 August 2017 Hazelcast IMDG 4.0 September 2017
© 2017 Hazelcast Inc. Confidential & Proprietary 73 New Hazelcast IMDG 3.8 Features Features Description Fast Aggregations Based on Queries Make aggregations significantly faster by moving in-built aggregations to be on top of the query infrastructure of Hazelcast IMDG. Another goal is to expose a simplified Aggregations API. User Code Deployment Ability to access and use user classes without having them on the classpath. Projection for Queries Provide projection in queries to allow returning specific fields of an entry, not the whole entry. This will cut the network and serialisation overhead and increase the throughput. RingBuffer Store Implementation of the storage mechanism for the RingBuffer, analogous to the existing QueueStore mechanism. It will also allow users of the ReliableTopic to replay messages older than the messages in the underlying RingBuffer. Continuous Query The previously Enterprise feature Continuous Query Cache becomes open source with 3.8.
© 2017 Hazelcast Inc. Confidential & Proprietary 74 New Hazelcast IMDG 3.8 Features Features Description Unified Near Cache Implementation for IMap and JCache A technical debt to have just 2 Near-Cache implementations: OnHeapNearCache and HDNearCache. These 2 implementations should be used by both IMap and JCache as well as member and client. High Performance Near Cache Several Improvements on Near Cache including pre-loading near cache. Eventually Consistent Near Cache A guarantee of delivery of cache invalidations to the near cache. HyperLogLog Implementation of the HyperLogLog data structure on top of Hazelcast IMDG. It is a probabilistic structure used to estimate cardinality. Scheduled Executor Service Distributed version of java.util.concurrent.ScheduledExecutorService. Cluster Quorum for Queue and Lock Quorum support for Hazelcast IMDG Queue and Lock.
© 2017 Hazelcast Inc. Confidential & Proprietary 75 New Hazelcast IMDG 3.8 Enterprise Features Features Description Rolling Member Upgrades for Minor Releases Extending current rolling upgrade capability of Hazelcast IMDG (client rolling upgradeability, patch version rolling member upgrades) to allow heterogeneous versions of Hazelcast IMDG members to be rolling upgraded without service interruption for minor version changes. Copying HRS Data from a Source to a Target Cluster Ability to copy the Hot Restart store files from a source environment to a target environment (for example, product to test environment). This feature also applies to running sources. One-Off WAN Sync with Dynamically Added Endpoint Ability to copy all Maps’ content to a target cluster dynamically in runtime. SSL Performance Improvements (Java) Large performance increase in SSL when using the Java client. LDAP Integration for Management Center Authentication and authorizarion with LDAP in Management Center. Many Management Center Enhancements More metrics, stats, and improved usability and monitoring.
© 2017 Hazelcast Inc. Confidential & Proprietary 76 New Hazelcast IMDG 3.8 Clients Features Features Description Mutual Auth with TLS Certificates Allow using client certificates to identify the client to the server. Simulator Integration for Hazelcast .NET Client Integration of the .NET client into Hazelcast Simulator. Node.js Client Enhancements RingBuffer, Topic, and Near Cache Support for Node.js client. SSL Support for Node.js Client SSL support for Hazelcast Node.js Client to secure the connection between the client and the members. Near Cache Support for C++ Client Configurable near cache support for IMap in the Hazelcast C++ client. AWS Discovery for C++ Client AWS cluster discovery implementation for Hazelcast C++ Client.
© 2017 Hazelcast Inc. Confidential & Proprietary 77 New Hazelcast IMDG 3.8 Cloud & Plugins Features Features Description Pivotal® CloudFoundry Integration Hazelcast IMDG on Pivotal CloudFoundry as a Tile. IBM DynaCache API Support A WebSphere DynamicCache provider, which could be used either by traditional WebSphere or WebSphere Liberty Profile applications. Discovery SPI based AWS Module Discovery SPI compatible hazelcast-aws module. AWS users would take advantage of ZONE_AWARE (availability zone aware partition groups) without using Apache jclouds. Striim Hot Cache Adapter HotCache functionality with Striim based on Change Data Capture ("CDC") of underlying database updates. Hibernate 5.2 Support Improvements in distributed L2 cache for Hibernate to support Hibernate 5.2.
© 2017 Hazelcast Inc. Confidential & Proprietary 78 Hazelcast 3.9 Roadmap August 2017
© 2017 Hazelcast Inc. Confidential & Proprietary 79 3.9 Features Features Description Near Cache: Key Store-by- Reference Option Key store-by-reference option for IMap and Hazelcast JCache Near Cache. Dynamic Creation of Specificly Configured Distributed Objects Ability to define new data structures with custom configuration on the fly. Gigantic Cache Migration Enhancements Improvements in partition migration process to allow better gigantic cache migrations. Lazy Initiation for Java Client An async client factory to connect lazily and in a non-blocking manner. Client Support for User Code Deployment Dynamic Classloading from clients for User Code Deployment. Query Async API Async query methods on the IMap
© 2017 Hazelcast Inc. Confidential & Proprietary 80 3.9 – Security Features Features Description Mutual TLS Auth Java Client Allow a member to authenticate a client using a certificate installed in the client Mutual TLS Auth C# Client Allow a member to authenticate a client using a certificate installed in the client Management Center TLS TLS from members to Man Center. TLS from browser. TLS for REST endpoint TLS for REST endpoint on Members General Security Hardening Implement things like password blacklisting, password complexity requirements, toggling off of non-secure mechanisms when secure version is enabled and so on Fast SSL/TLS Make SSL/TLS almost as fast as unencrypted.
© 2017 Hazelcast Inc. Confidential & Proprietary 81 Fast SSL/TLS
© 2017 Hazelcast Inc. Confidential & Proprietary 82 New Hazelcast IMDG 3.9 Enterprise Features Features Description Dynamic Addition of WAN Endpoints Ability to add WAN Targets to a running cluster. NearCache Stats in Management Center Provision of showing data inside Near-Cache of client or member, similar to IMap/Cache in Management Center. Search Indexes in HD Memory Support for sorted and unsorted indexes in Hazelcast HD Memory Store.
© 2017 Hazelcast Inc. Confidential & Proprietary 83 New Hazelcast IMDG 3.9 Clients Features Features Description Simulator Integrations of C++, Python, and Node.js Clients Integration of C++, Python, and Node.js clients into Hazelcast Simulator. SSL Performance Improvements for .NET Client Large performance increase in SSL when using the .NET Client. Projections for .NET Client Add projections (returning attributes rather than full values) to .NET. Paging Predicate for .NET Client Support for paging predicates in Hazelcast .NET Client. ExecutorService for C++ Client Ability to use the distributed ExecutorService from Hazelcast C++ Client. Near Cache for C++ Client Add near cache to C++ client TLS for C++ Client Add TLS encryption for C++ Clients communicating with members
© 2017 Hazelcast Inc. Confidential & Proprietary 84 New Hazelcast IMDG 3.9 Cloud & Plugins Features Features Description Google Cloud Platform Integration SPI Plugin for GCP for Cloud Discovery with a support for partition groups. AWS support for C++ Client We have comprehensive cloud support for Java. Adding AWS to C++. Stats for Hibernate LocalRegionCache Ability to provide statistics for Hibernate LocalRegionCache. Web Sessions for WebSphere Liberty Web sessions replication plugin for WebSphere Liberty.
© 2017 Hazelcast Inc. Confidential & Proprietary 85 New Hazelcast IMDG 4.0 Features Features Description Java 8 Language Level Lift the language level for Hazelcast members and their API to Java 8. Add a new Java 8 client. Continue to maintain the Java 6 compatible client. This will allow us to take advantage of Java 8 language features. Java 9 Modularisation Java 9 modularisation for Hazelcast IMDG, which introduces Jigsaw - a new modular systems with a strict enforcement of module boundaries. Proper Separation of Public and Private APIs Re-design of the whole private/public API to keep as much private as possible. Public and Stable SPI Make SPI a second public and stable API. This work includes many refactorings, client protocol enhancements, documentation efforts, and reference implementations. EntryProcessor Contract Formalisation Optimization of the EntryProcessor which would bring the IMap behaviour inline with the JCache behaviour. Unification of Client / Member Behaviour in Exceptional States Equalization of exceptions in client-side and member-side so that clients can throw the same exceptions as server whenever possible. This work also includes a documentation effort on exceptional cases for each method.
What’s Hazelcast Jet? • General purpose distributed data processing framework • Based on Direct Acyclic Graph to model data flow • Built on top of Hazelcast • Comparable to Apache Spark or Apache Flink 8 7
DAG 8 8
Job Execution 8 9
9 0 Hazelcast Services
Service Offerings Hazelcast (Apache Licensed) • Professional Subscription – 24x7 support* Hazelcast Enterprise Support • Available with Hazelcast Enterprise software subscription - 24x7 support* Additional Services • Development Support Subscription – 8x5 support* • Simulator TCK • Training • Expert Consulting • Development Partner Program * All subscriptions include Management Center
© 2016 Hazelcast Inc. Confidential & Proprietary 92 Support Subscriptions What’s Included 100% SUCCESS RATE ON CUSTOMER ISSUES: “As usual, the response was timely beyond expectations, and very good technical content returned. Exemplary support, hard to find in any company…” - Fortune 100 Financial Services Customer ENTERPRISE HD ENTERPRISE PROFESSIONAL OPEN SOURCE SUPPORT WINDOW 24/7 24/7 24/7 RESPONSE TIME FOR CRITIAL ISSUES 1 Hour 1 Hour 2 Hours SUPPORTED SOFTWARE Hazelcast & Hazelcast Enterprise Hazelcast & Hazelcast Enterprise Hazelcast SUPPORT CONTACTS 4 4 2 SUPPORT CHANNELS Email, IM & Phone Email, IM & Phone Email, IM & Phone PATCH LEVEL FIXES REMOTE MEETINGS (via GoToMeeting) CODE REVIEW (with a Senior Solutions Architect) 2 Hours 2 Hours 2 Hours QUARTERLY REVIEW OF FEATURE REQUES* QUARTERLY REVIEW OF HAZELCAST ROADMAP*
© 2016 Hazelcast Inc. Confidential & Proprietary 9 3 Best In Class Support Support from the Engineers who wrote the code SLA Driven – 100% attainment of support response time Follow the Sun Portal, Email and Phone access Go Red, Go Green. Reproduction of issues on Simulator. Proof of fix on Simulator. Periodic Technical Reviews Meet your production schedule and corporate compliance requirements Ensure the success of your development team with training and best practices
9 4 Hazelcast Support Coverage India Turkey London U.S.
© 2016 Hazelcast Inc. Confidential & Proprietary 95 Support Testimonials
© 2016 Hazelcast Inc. Confidential & Proprietary 96 Release Lifecycle • Regular Feature release each 4-5 months, e.g. 3.3, 3.4, 3.5 • Maintenance release approximately each month with bug fixes based on the current feature release, e.g. 3.4.1 • For older versions, patch releases made available to fix issues • Release End of Life per support contract
Thank you

More Related Content

PPTX
Hazelcast Jet - January 08, 2018
byRahul Gupta
 
PDF
Openstack
byRAKESH SHARMA
 
PPTX
Hazelcast Jet v0.4 - August 9, 2017
byRahul Gupta
 
PDF
JCConf 2016 - Cloud Computing Applications - Hazelcast, Spark and Ignite
byJoseph Kuo
 
PPTX
Tez Data Processing over Yarn
byInMobi Technology
 
PPT
Deep Learning with DL4J on Apache Spark: Yeah it's Cool, but are You Doing it...
byDataWorks Summit
 
PPTX
February 2014 HUG : Hive On Tez
byYahoo Developer Network
 
PPTX
Apache Tez -- A modern processing engine
bybigdatagurus_meetup
 
Hazelcast Jet - January 08, 2018
byRahul Gupta
 
Openstack
byRAKESH SHARMA
 
Hazelcast Jet v0.4 - August 9, 2017
byRahul Gupta
 
JCConf 2016 - Cloud Computing Applications - Hazelcast, Spark and Ignite
byJoseph Kuo
 
Tez Data Processing over Yarn
byInMobi Technology
 
Deep Learning with DL4J on Apache Spark: Yeah it's Cool, but are You Doing it...
byDataWorks Summit
 
February 2014 HUG : Hive On Tez
byYahoo Developer Network
 
Apache Tez -- A modern processing engine
bybigdatagurus_meetup
 

What's hot

PPTX
Geek Nights Hong Kong
byRahul Gupta
 
PDF
Machine Learning for Capacity Management
byEDB
 
PPTX
Pig on Tez: Low Latency Data Processing with Big Data
byDataWorks Summit
 
PPTX
Tuning up with Apache Tez
byGal Vinograd
 
PPTX
What's the Hadoop-la about Kubernetes?
byDataWorks Summit
 
PDF
Red Hat Ceph Storage: Past, Present and Future
byRed_Hat_Storage
 
PPTX
Tez big datacamp-la-bikas_saha
byData Con LA
 
PPTX
Apache Tez - Accelerating Hadoop Data Processing
byhitesh1892
 
PPTX
20150314 sahara intro and the future plan for open stack meetup
byWei Ting Chen
 
PPTX
YARN Containerized Services: Fading The Lines Between On-Prem And Cloud
byDataWorks Summit
 
PDF
Hazelcast 3.6 Roadmap Preview
byHazelcast
 
PPTX
OLTP+OLAP=HTAP
byEDB
 
PPTX
Hazelcast Deep Dive (Paris JUG-2)
byEmrah Kocaman
 
PPTX
20151027 sahara + manila final
byWei Ting Chen
 
PDF
The Future of Computing is Distributed
byAlluxio, Inc.
 
PPTX
Apache Tez : Accelerating Hadoop Query Processing
byBikas Saha
 
PPTX
Hive+Tez: A performance deep dive
byt3rmin4t0r
 
ODP
Sahara presentation latest - Codemotion Rome 2015
byCodemotion
 
PPTX
Hazelcast Essentials
byRahul Gupta
 
PDF
BlueData and Hortonworks Data Platform (HDP)
byBlueData, Inc.
 
Geek Nights Hong Kong
byRahul Gupta
 
Machine Learning for Capacity Management
byEDB
 
Pig on Tez: Low Latency Data Processing with Big Data
byDataWorks Summit
 
Tuning up with Apache Tez
byGal Vinograd
 
What's the Hadoop-la about Kubernetes?
byDataWorks Summit
 
Red Hat Ceph Storage: Past, Present and Future
byRed_Hat_Storage
 
Tez big datacamp-la-bikas_saha
byData Con LA
 
Apache Tez - Accelerating Hadoop Data Processing
byhitesh1892
 
20150314 sahara intro and the future plan for open stack meetup
byWei Ting Chen
 
YARN Containerized Services: Fading The Lines Between On-Prem And Cloud
byDataWorks Summit
 
Hazelcast 3.6 Roadmap Preview
byHazelcast
 
OLTP+OLAP=HTAP
byEDB
 
Hazelcast Deep Dive (Paris JUG-2)
byEmrah Kocaman
 
20151027 sahara + manila final
byWei Ting Chen
 
The Future of Computing is Distributed
byAlluxio, Inc.
 
Apache Tez : Accelerating Hadoop Query Processing
byBikas Saha
 
Hive+Tez: A performance deep dive
byt3rmin4t0r
 
Sahara presentation latest - Codemotion Rome 2015
byCodemotion
 
Hazelcast Essentials
byRahul Gupta
 
BlueData and Hortonworks Data Platform (HDP)
byBlueData, Inc.
 

Similar to Distributed caching-computing v3.8

PPTX
Think Distributed: The Hazelcast Way
byRahul Gupta
 
PPTX
Distributed caching and computing v3.7
byRahul Gupta
 
PDF
Building scalable applications with hazelcast
byFuad Malikov
 
PDF
Building scalable applications with hazelcast
byFuad Malikov
 
PPT
Hazelcast
byJeevesh Pandey
 
PDF
Clustering your Application with Hazelcast
byHazelcast
 
PDF
Hazelcast for Terracotta Users
byHazelcast
 
PDF
Hazelcast 101
byEmrah Kocaman
 
PDF
In-memory No SQL- GIDS2014
byHazelcast
 
PDF
Easy Scaling with Open Source Data Structures, by Talip Ozturk
byZeroTurnaround
 
PPTX
From cache to in-memory data grid. Introduction to Hazelcast.
byTaras Matyashovsky
 
PDF
Hazelcast Introduction
byCodeOps Technologies LLP
 
PPTX
Hazelcast sunum
bySoftware Infrastructure
 
PPTX
IMC Summit 2016 Breakout - Greg Luck - How to Speed Up Your Application Using...
byIn-Memory Computing Summit
 
PDF
Distributed applications using Hazelcast
byTaras Matyashovsky
 
PPTX
Hazelcast For Beginners (Paris JUG-1)
byEmrah Kocaman
 
PDF
Devoxx uk 2014 High performance in-memory Java with open source
byDavid Brimley
 
PDF
Devoxx 2013 - Hazelcast
byHazelcast
 
PDF
Caching and JCache with Greg Luck 18.02.16
byComsysto Reply GmbH
 
PPTX
Hazelcast and MongoDB at Cloud CMS
byuzquiano
 
Think Distributed: The Hazelcast Way
byRahul Gupta
 
Distributed caching and computing v3.7
byRahul Gupta
 
Building scalable applications with hazelcast
byFuad Malikov
 
Building scalable applications with hazelcast
byFuad Malikov
 
Hazelcast
byJeevesh Pandey
 
Clustering your Application with Hazelcast
byHazelcast
 
Hazelcast for Terracotta Users
byHazelcast
 
Hazelcast 101
byEmrah Kocaman
 
In-memory No SQL- GIDS2014
byHazelcast
 
Easy Scaling with Open Source Data Structures, by Talip Ozturk
byZeroTurnaround
 
From cache to in-memory data grid. Introduction to Hazelcast.
byTaras Matyashovsky
 
Hazelcast Introduction
byCodeOps Technologies LLP
 
Hazelcast sunum
bySoftware Infrastructure
 
IMC Summit 2016 Breakout - Greg Luck - How to Speed Up Your Application Using...
byIn-Memory Computing Summit
 
Distributed applications using Hazelcast
byTaras Matyashovsky
 
Hazelcast For Beginners (Paris JUG-1)
byEmrah Kocaman
 
Devoxx uk 2014 High performance in-memory Java with open source
byDavid Brimley
 
Devoxx 2013 - Hazelcast
byHazelcast
 
Caching and JCache with Greg Luck 18.02.16
byComsysto Reply GmbH
 
Hazelcast and MongoDB at Cloud CMS
byuzquiano
 

Recently uploaded

PPTX
Top Trends in Kubernetes Security: TalosCon 2025
byCheryl Hung
 
PDF
Getting the Best of TrueDEM – October News & Updates
bypanagenda
 
PDF
A fool with a tool is still a fool - Plone Conference 2025
byAndreas Jung
 
PPTX
Session 5 - Specialized AI Associate Series: Build a Document Understanding ...
byDianaGray10
 
PDF
The invisible key: Securing the new attack vector of OAuth tokens
byGianluca Varisco
 
PDF
Cassandra vs. ScyllaDB: Evolutionary Differences
byScyllaDB
 
PPTX
Agentic AI Solutions and Services | Aeologic
byankitaeologic
 
PDF
TrustArc Webinar - Migration to TrustArc: What Your Journey Will Look Like
byTrustArc
 
PDF
Meta and Apple close to settling EU cases.pdf
byLUMINATIVE MEDIA/PROJECT COUNSEL MEDIA GROUP
 
PDF
Unlocking Full-Stack Observability for AIOps: A Precisely Ironstream for Big ...
byPrecisely
 
PDF
Using Copilot with Microsoft Office Apps
byDeb Walther
 
PPTX
UiPath Automation Suite Installation (Hands-On) [2/3]
bysuhanisingh58689
 
PDF
How to Measure Microsoft 365 Copilot Success and Manage AI Risk: Advanced Str...
byNikki Chapple
 
PDF
Ethical Initiatives in AI and accountability.pdf
byShiwani Gupta
 
PDF
NSSHOEU-J 4x35mm² Cable Specification.pdf
byAnhui Feichun Special Cable Co., Ltd.
 
PDF
UnityNet Digital Sovereignty Checklist 2025-10-13
byLHelferty
 
PDF
Session 4 - Specialized AI Associate Series: UiPath Document Understanding O...
byDianaGray10
 
PDF
Planetek Italia Corporate Profile brochure
byPlanetek Italia Srl
 
PDF
Ask Me Anything About AI Assist: Practical Answers for Real Workflows
bySafe Software
 
PDF
Fairness and Bias in AI Ethics and Explainability
byShiwani Gupta
 
Top Trends in Kubernetes Security: TalosCon 2025
byCheryl Hung
 
Getting the Best of TrueDEM – October News & Updates
bypanagenda
 
A fool with a tool is still a fool - Plone Conference 2025
byAndreas Jung
 
Session 5 - Specialized AI Associate Series: Build a Document Understanding ...
byDianaGray10
 
The invisible key: Securing the new attack vector of OAuth tokens
byGianluca Varisco
 
Cassandra vs. ScyllaDB: Evolutionary Differences
byScyllaDB
 
Agentic AI Solutions and Services | Aeologic
byankitaeologic
 
TrustArc Webinar - Migration to TrustArc: What Your Journey Will Look Like
byTrustArc
 
Meta and Apple close to settling EU cases.pdf
byLUMINATIVE MEDIA/PROJECT COUNSEL MEDIA GROUP
 
Unlocking Full-Stack Observability for AIOps: A Precisely Ironstream for Big ...
byPrecisely
 
Using Copilot with Microsoft Office Apps
byDeb Walther
 
UiPath Automation Suite Installation (Hands-On) [2/3]
bysuhanisingh58689
 
How to Measure Microsoft 365 Copilot Success and Manage AI Risk: Advanced Str...
byNikki Chapple
 
Ethical Initiatives in AI and accountability.pdf
byShiwani Gupta
 
NSSHOEU-J 4x35mm² Cable Specification.pdf
byAnhui Feichun Special Cable Co., Ltd.
 
UnityNet Digital Sovereignty Checklist 2025-10-13
byLHelferty
 
Session 4 - Specialized AI Associate Series: UiPath Document Understanding O...
byDianaGray10
 
Planetek Italia Corporate Profile brochure
byPlanetek Italia Srl
 
Ask Me Anything About AI Assist: Practical Answers for Real Workflows
bySafe Software
 
Fairness and Bias in AI Ethics and Explainability
byShiwani Gupta
 

Distributed caching-computing v3.8

  • 1.
    Thinking Distributed: TheHazelcast Way RAHUL GUPTA SR SOLUTIONS ARCHITECT
  • 2.
    About Me Senior SolutionsArchitect Worked with Terracotta In-memory Distributed Systems since 2009 Java Programmer since 1998 rahul@hazelcast.com Rahul Gupta follow me @wildnez
  • 3.
    © 2017 HazelcastInc. Confidential & Proprietary 3 Hazelcast Reborn
  • 4.
  • 5.
    Why Hazelcast? • Scale-outComputing enables cluster capacity to be increased or decreased on-demand • Resilience with automatic recovery from member failures without losing data while minimizing performance impact on running applications • Programming Model provides a way for developers to easily program a cluster application as if it is a single process • Fast Application Performance enables very large data sets to be held in main memory for real-time performance
  • 6.
    © 2017 HazelcastInc. Confidential & Proprietary 6 Low-Level Services API Node Engine (Threads, Instances, Eventing, Wait/Notify, Invocation) Partition Management (Members, Lite Members, Master Partition, Replicas, Migrations, Partition Groups, Partition Aware) Cluster Management with Cloud Discovery SPI (AWS, Azure, Consul, etcd, Heroku, IP List, Apache jclouds, Kubernetes, Multicast, Zookeeper) Networking (IPv4, IPv6) On-Heap Store High-Density Memory Store (Intel, Sparc) Serialization (Serializable, Externalizable, DataSerializable, IdentifiedDataSerializable, Portable, Custom) Near Cache HD C++ C#/.NET Python Node.js Scala Memcached Clojure Open Client Network Protocol (Backward & Forward Compatibility, Binary Protocol) REST Java JVM (JDK: 6, 7, 8, Vendors: Oracle JDK, Open JDK, IBM JDK, Azul Zing & Zulu) Hardware/VM/Container (On Premise, AWS, Azure, Docker, Pivotal Cloud Foundry, OpenShift, VMware) Executor Service SQL Query MapReduce AggregationPredicate Entry Processor Map MultiMap Replicated Map Web Sessions (Tomcat/Jetty/Generic) Continuous QueryHD HD Open Source Enterprise Edition Hazelcast IMDG Enterprise HD Edition HD-Enabled Feature WAN (Socket,SolaceSystems:One-way,Multiway,InitnewDataCenter,DRDataCenterRecovery) ManagementCenter (JMX/REST) SecuritySuite (Connection,Encryption,Authentication,Authorization,JAASLoginModule,SocketInterceptor) Engine Clients HD HD HD Queue Topic Ring Buffer Atomic Long Hibernate 2nd Level Cache HD JCache Set Lock/ Sem. HD java.util.concurrent UUID Hot Restart Store (SSD, HDD) HD Storage Operating System (Linux, Oracle Solaris, Windows, AIX, Unix) Operating Env APIs Hyper LogLog RollingUpgrades (RollingClientUpgrades,RollingMemberUpgrades,NoDowntime) HazelcastStriimHotCache (SyncUpdatesfromOracleDB,MSSQLServer,MySQLandNonStopDB) Operations Hazelcast Solution
  • 7.
    Ecosystem Traction Dozens ofCommercial and Open Source Projects Embed Hazelcast
  • 8.
    01001 10101 01010 In Memory Data Computing InMemory Data Messaging++In Memory Data Storage In Memory Data Grid
  • 9.
    Business Systems A BC RDBMS Mainframe MongoDB NoSQL REST Scale Hazelcast In-Memory Caching
  • 10.
    java.util.concurrent.ConcurrentMap import java.util.concurrent.ConcurrentHashMap; import java.util.concurrent.ConcurrentMap; publicstatic void main(String[] args) { ConcurrentMap<Integer, String> map = new ConcurrentHashMap<>(); map.put(1, "Paris"); map.put(2, "London"); map.put(3, "San Francisco"); String oldValue = map.remove(2); }
  • 11.
    Distributed Map import java.util.concurrent.ConcurrentMap; importcom.hazelcast.core.Hazelcast; import com.hazelcast.core.HazelcastInstance; public static void main(String[] args) { HazelcastInstance h = Hazelcast.newHazelcastInstance(); ConcurrentMap<Integer, String> map = h.getMap("myMap"); map.put(1, "Paris"); map.put(2, "London"); map.put(3, "San Francisco"); String oldValue = map.remove(2); }
  • 12.
  • 13.
    Persistence API public classMapStorage implements MapStore<String, User>, MapLoader<String, User> { // Some methods missing ... @Override public User load(String key) { return loadValueDB(key); } @Override public Set<String> loadAllKeys() { return loadKeysDB(); } @Override public void delete(String key) { deleteDB(key); } @Override public void store(String key, User value) { storeToDatabase(key, value); } } <map name="users"> <map-store enabled="true"> <class-name>com.hazelcast.example.MapStorage</class-name> <write-delay-seconds>0</write-delay-seconds> </map-store> </map>
  • 14.
    JCache API // Retrievethe CachingProvider which is automatically baced by // the chosen Hazelcast server or client provider CachingProvider cachingProvider = Caching.getCachingProvider(); // Create a CacheManager CacheManager cacheManager = cachingProvider.getCacheManager(); // Cache<String, String> cache = cacheManager // .getCache( name, String.class, String.class ); // Create a simple but typesafe configuration for the cache CompleteConfiguration<String, String> config = new MutableConfiguration<String, String>() .setTypes( String.class, String.class );
  • 15.
    JCache API // Createand get the cache Cache<String, String> cache = cacheManager .createCache( "example", config ); // Alternatively to request an already existing cache Cache<String, String> cache = cacheManager .getCache( name, String.class, String.class ); // Put a value into the cache cache.put( "world", "Hello World" ); // Retrieve the value again from the cache String value = cache.get( "world" ); System.out.println( value );
  • 16.
    Eviction • Unless deleted,entries remain in the map. • Use eviction policies to prevent OOM situations. • Eviction Triggers run LFU or LRU. time-to-live max-idle-seconds max-size (PER_NODE(number of entries), PER_PARTITION(number of entries), USED_HEAP_SIZE(mb), USED_HEAP_PERCENTAGE(mb)) • Setting eviction-percentage removes that % of entries when eviction is triggered.
  • 17.
    Features Description MultiMap Storemultiple values against one Key Replicated Map Cluster wide replication, all entries on all nodes. Good for read heavy use cases Near Cache Map Entries on Client/Application local memory RingBuffer Stores data in a ring-like structure, like a circular array with given capacity More Distributed Structures
  • 18.
    Java Collection API:Map, List, Set, Queue Jcache High Density Memory Store Hibernate 2nd Level Cache Web Session Replication: Tomcat, Jetty Predicate API: Indexes, SQL Query Persistence: Map/Queue Store & Loader. Write Behind/Through Spring Compliance Transactions: Local & XA WAN & DR Replication IM Data Store (Caching) Features
  • 19.
    Java: Will itmake the cut? Garbage Collection limits heap usage. G1 and Balanced aim for <100ms at 10GB. Unused Memory 64GB 4GB 4s Heap Java Apps Memory Bound GC Pause Time Available Memory GC Off-Heap Storage No low-level CPU access Java is challenged as an infrastructure language despite its newly popular usage for this Heap
  • 20.
    GC Pause 64 Gb JVM 4Gb JVM 1 4 Gb JVM 2 4 Gb JVM 3 4 Gb JVM 4 4 Gb JVM 5 4 Gb JVM 16 GC Pause 0 GB 64 GB 64 Gb in-memory data time Standard Impediments of Caching
  • 21.
    GC Pause 64 Gb JVM 4Gb JVM 1 4 Gb JVM 2 4 Gb JVM 3 4 Gb JVM 4 4 Gb JVM 5 4 Gb JVM 16 GC Pause 0 Gb 64 Gb 64 Gb in-memory data time 64 Gb HD 2 Gb JVM GC Pause HD Memory Caching with HD Memory
  • 22.
    On-Heap Memory Store (ObjectsStored as Objects) High-Density Memory Store (Objects Serialized and Stored as Bytes) s.m.Unsafe s.m.Unsafe 2-4GB (Limited by Garbage Collection) 0-1TB (Limited by Machine RAM) Memory Stores •Member •Client (Near Cache) RAM in JVM Process APIs JCache (ICache) Map (IMap) HD Memory
  • 23.
    On Heap Vs.High-Density Memory Management On Heap Memory HD Memory 0 MB HD 3.3 GB 3.9 GB Heap Storage 0.6 GB 9 (4900 ms) Major GC 0 (0 ms) 31 (4200 ms) Minor GC 356 (349 ms) Node Used Heap Total Heap Max. Heap Heap Usage Percentage Used Heap: 0.2 GB 192.168.1.10:5701 57 MB 229 MB 910 MB 6.28% Memory Utilization Home Offheap-test Node Used Heap: 3.9 GB 192.168.1.10:5701 3933 MB 4658MB 4653MB 84.45% Memory Utilization Home Used Heap Total Heap Max. Heap Heap Usage Percentage Example: On Heap Memory Example: HD Memory
  • 24.
    Hazelcast Servers Hazelcast Server JVM[Memory] A B C Business Logic Data Data Data CE = Compute Engine Result Business / Processing Logic Result TCP / IP Client Client Distributed Computing
  • 25.
    IM Distributed ComputingFeature HD Cache Dist. Compute Dist. Message Java Concurrency API (Lock, Semaphore, AtomicLong, AtomicReference, Executor Service, Blocking Queue) Entry and Item Listeners Entry Processor Aggregators Map/Reduce Data Affinity Continues Query Map Interceptors Delta Update
  • 26.
    Executor Service API publicinterface com.hazelcast.core.IExecutorService extends java.util.concurrent.ExecutorService HazelcastInstance hz = getHazelcastInstance(); //java.util.concurrent.ExecutorService implementation IExecutorService es = hz.getExecutorService("name"); es.executeOnAllMembers(buildRunnable()); es.executeOnKeyOwner(buildRunnable(), "Peter"); es.execute(buildRunnable()); Map<..> futures = es.submitToAllMembers(buildCallable()); Future<..> future = es.submitToKeyOwner(buildCallable(), "Peter"); es.submitToAllMembers(buildCallable(), buildCallback()); es.submitToKeyOwner(buildCallable(), "Peter", buildCallback());
  • 27.
  • 28.
    Lock API HazelcastInstance hz= getHazelcastInstance(); // Distributed Reentrant L​ock l​ock = hz.getLock("myLock"); l​ock.lock(); try { // Do something } finally { l​ock.unlock(); } Distributed Lock
  • 29.
  • 30.
  • 31.
    Map/Reduce API HazelcastInstance hz= getHazelcastInstance(); Map users = hz.getMap("users"); JobTracker tracker = hz.getJobTracker("default"); KeyValueSource source = KeyValueSource.fromMap(users); Job job = tracker.newJob(source); ICompleteFuture future = job.mapper(new MyMapper()) .reducer(new MyReducer()) .submit(); Map result = future.get();
  • 32.
    Aggregations API HazelcastInstance hz= getHazelcastInstance(); Map users = hz.getMap("users"); int sum = users.aggregate( Supplier.all((user) -> user.getSalary()), Aggregations.longSum() );
  • 33.
    Hazelcast Distributed Topic Bus Hazelcast Topic Hazelcast Node1 Hazelcast Node 2 Hazelcast Node 3 MSG Subscribes Delivers Subscribes Delivers Distributed Messaging
  • 34.
    Queue Topic (Pub/Sub) Event Listeners RingBuffers IM Distributed Messaging Features HD Cache Dist. Compute Dist. Message
  • 35.
    Queue API interface com.hazelcast.core.IQueue<E> extendsjava.util.concurrent.BlockingQueue HazelcastInstance hz = getHazelcastInstance(); //java.util.concurrent.BlockingQueue implementation IQueue<Task> queue = hz.getQueue("tasks"); queue.offer(newTask()); queue.offer(newTask(), 500, TimeUnit.MILLISECONDS); Task task = queue.poll(); Task task = queue.poll(100, TimeUnit.MILLISECONDS); Task task = queue.take();
  • 36.
    Topic API public classExample implements MessageListener<String> { public void sendMessage { HazelcastInstance hz = getHazelcastInstance(); ITopic<String> topic = hz.getTopic("topic"); topic.addMessageListener(this); topic.publish("Hello World"); } @Override public void onMessage(Message<String> message) { System.out.println("Got message: " + message.getMessageObject()); } }
  • 38.
    Data Distribution andResilience 38
  • 39.
    Distributed Maps Fixed numberof partitions (default 271) Each key falls into a partition partitionId = hash(keyData)%PARTITION_COUNT Partition ownerships are reassigned upon membership change A B C
  • 40.
  • 41.
  • 42.
  • 43.
  • 44.
  • 45.
  • 46.
  • 47.
  • 48.
    Data Safety onNode Failure 48
  • 49.
  • 50.
  • 51.
  • 52.
  • 53.
  • 54.
  • 55.
  • 56.
  • 57.
  • 58.
  • 59.
    © 2017 HazelcastInc. Confidential & Proprietary 59 Topologies Deployment Architectures
  • 60.
    © 2017 HazelcastInc. Confidential & Proprietary 60 Embedded Mode Great for microservices, OEM and ops simplification IMDG Application Java API IMDG Application Java API IMDG Application Java API
  • 61.
    © 2017 HazelcastInc. Confidential & Proprietary 61 Embedded Mode Embedded model, for example in a J2EE container Service and storage in the one JVM Easy for getting started With three nodes, 1/3rd of the data is local, good read performance -> This benefit diminishes as you scale up -> With ten nodes, only 1/10th of the data is local If the application needs to go offline, so does the data
  • 62.
    © 2017 HazelcastInc. Confidential & Proprietary 62 Client-Server Mode Great for scale-up or scale-out deployments with cluster lifecycle decoupled from app servers Use for non-Java clients Node.js API Application Node.js API Application Node.js API Application Node.js API ApplicationJava API Application .Net API Application C++ API Application C++ API Application C++ API Application C++ API Application .Net API Application .Net API Application .Net API ApplicationJava API Application Java API Application IMDG IMDG IMDG
  • 63.
    © 2017 HazelcastInc. Confidential & Proprietary 63 Client-Server Mode Follows the traditional Client-Server approach Clients do not change cluster topology Allows you to segregate service from storage Smart clients connect to all cluster members. Operations are routed directly to a member holding the necessary data. -> One network hop Other clients connect to a single cluster member and use it as a proxy to issue requests. -> Two network hops if the proxy doesn’t have the data needed
  • 64.
    © 2017 HazelcastInc. Confidential & Proprietary 64 Pushing Into New Language Ecosystems Hazelcast.org/Plugins • Fully featured • Hazelcast supported • Open Source • Common Hazelcast Feature Set • Hazelcast supported • Open Source • High Performance • Hazelcast supported • Open Source • Client and Member APIs • Scala 2.11 support • Hazelcast Supported • Open Source • Community supported • Open Source • Fully Featured • Hazelcast supported • Open Source • Industry Leading Performance • High-Density Memory Store Near Cache • Hazelcast Supported • Open Source Feature Matrix: https://hazelcast.org/clients-languages
  • 65.
    © 2017 HazelcastInc. Confidential & Proprietary 65 Hazelcast IMDG Microservices Deployment Options Microservices deployed embedded with isolated Hazelcast IMDG clusters Maximum service isolation Microservices deployed client server with a shared Hazelcast IMDG cluster Transition to microservices MS 1 IMDG IMDG IMDG MS 1 MS 1 .NET Client MS 1 IMDG IMDG IMDG IMDG .NET Client MS 1 Node.js Client MS 2 Node.js Client MS 2 C++ Client MS 3 Java Client MS 4 Java Client MS 4 MS 2 IMDG IMDG IMDG MS 2 MS 2 MS 3 IMDG IMDG IMDG MS 3 MS 3
  • 66.
    © 2017 HazelcastInc. Confidential & Proprietary 66 Hazelcast IMDG Microservices Deployment Options Microservices deployed client server with isolated Hazelcast cluster per service Maximum service isolation and non-Java clients .NET Client MS 1 IMDG IMDG IMDG .NET Client MS 1 .NET Client MS 1 .NET Client MS 1 .NET Client MS 1 Python Client MS 2 IMDG IMDG IMDG Python Client MS 2 Python Client MS 2 Python Client MS 2 Python Client MS 2
  • 67.
    Same Features –Embedded vs Client-Server // Creating a new Hazelcast node HazelcastInstance hz = Hazelcast.newHazelcastInstance(); // Getting a Map, Queue, Topic, ... Map map = hz.getMap("my-map"); Queue queue = hz.getQueue("my-queue"); ITopic topic = hz.getTopic("my-topic"); //Creating a Hazelcast Client HazelcastInstance client = HazelcastClient.newHazelcastClient(); // Shutting down the node hz.shutdown();
  • 68.
    © 2017 HazelcastInc. Confidential & Proprietary 68 Hazelcast IMDG – Any Cloud, Anytime Apache jClouds PaaS Environment Integrations Cloud & Cloud Discovery Plugins Deployment Options YARN
  • 69.
    © 2017 HazelcastInc. Confidential & Proprietary 69 Hazelcast IMDG WAN Replication Deployment Options Independent clusters, sharing selected data to maintain alignment One-way or two-way Choose which data to share, select by type, can de-select records by attributes One-way  keep DR site populated and ready-to-go Two-way  worldwide operations, access your nearest data copy API Applications API Applications API Applications IMDG IMDG IMDG IMDG IMDG IMDG IMDG
  • 70.
  • 71.
    © 2017 HazelcastInc. Confidential & Proprietary 71 Hazelcast IMDG High Level Roadmap Hi-Density Caching In-Memory Data Grid 2015 2016 JCache | Scalability | HD Memory | Hot Restart | Pre-Load Clients Cloud Discovery SPI | Azure Cloud Foundry | AWS j.u.c. | Low-Latency | Elasticity | Resiliency | Performance | Rolling Upgrades Open Cloud Platform 2017 j.u.s | Jet | Cascading | Streaming | Mesos | Reactive Expanded In-memory Computing 2018
  • 72.
    © 2017 HazelcastInc. Confidential & Proprietary 72 Recent Releases & Roadmap Release Date Hazelcast IMDG 3.7 August 2016 (Released) Hazelcast IMDG 3.8 February 2017 (Released) Hazelcast IMDG 3.9 August 2017 Hazelcast IMDG 4.0 September 2017
  • 73.
    © 2017 HazelcastInc. Confidential & Proprietary 73 New Hazelcast IMDG 3.8 Features Features Description Fast Aggregations Based on Queries Make aggregations significantly faster by moving in-built aggregations to be on top of the query infrastructure of Hazelcast IMDG. Another goal is to expose a simplified Aggregations API. User Code Deployment Ability to access and use user classes without having them on the classpath. Projection for Queries Provide projection in queries to allow returning specific fields of an entry, not the whole entry. This will cut the network and serialisation overhead and increase the throughput. RingBuffer Store Implementation of the storage mechanism for the RingBuffer, analogous to the existing QueueStore mechanism. It will also allow users of the ReliableTopic to replay messages older than the messages in the underlying RingBuffer. Continuous Query The previously Enterprise feature Continuous Query Cache becomes open source with 3.8.
  • 74.
    © 2017 HazelcastInc. Confidential & Proprietary 74 New Hazelcast IMDG 3.8 Features Features Description Unified Near Cache Implementation for IMap and JCache A technical debt to have just 2 Near-Cache implementations: OnHeapNearCache and HDNearCache. These 2 implementations should be used by both IMap and JCache as well as member and client. High Performance Near Cache Several Improvements on Near Cache including pre-loading near cache. Eventually Consistent Near Cache A guarantee of delivery of cache invalidations to the near cache. HyperLogLog Implementation of the HyperLogLog data structure on top of Hazelcast IMDG. It is a probabilistic structure used to estimate cardinality. Scheduled Executor Service Distributed version of java.util.concurrent.ScheduledExecutorService. Cluster Quorum for Queue and Lock Quorum support for Hazelcast IMDG Queue and Lock.
  • 75.
    © 2017 HazelcastInc. Confidential & Proprietary 75 New Hazelcast IMDG 3.8 Enterprise Features Features Description Rolling Member Upgrades for Minor Releases Extending current rolling upgrade capability of Hazelcast IMDG (client rolling upgradeability, patch version rolling member upgrades) to allow heterogeneous versions of Hazelcast IMDG members to be rolling upgraded without service interruption for minor version changes. Copying HRS Data from a Source to a Target Cluster Ability to copy the Hot Restart store files from a source environment to a target environment (for example, product to test environment). This feature also applies to running sources. One-Off WAN Sync with Dynamically Added Endpoint Ability to copy all Maps’ content to a target cluster dynamically in runtime. SSL Performance Improvements (Java) Large performance increase in SSL when using the Java client. LDAP Integration for Management Center Authentication and authorizarion with LDAP in Management Center. Many Management Center Enhancements More metrics, stats, and improved usability and monitoring.
  • 76.
    © 2017 HazelcastInc. Confidential & Proprietary 76 New Hazelcast IMDG 3.8 Clients Features Features Description Mutual Auth with TLS Certificates Allow using client certificates to identify the client to the server. Simulator Integration for Hazelcast .NET Client Integration of the .NET client into Hazelcast Simulator. Node.js Client Enhancements RingBuffer, Topic, and Near Cache Support for Node.js client. SSL Support for Node.js Client SSL support for Hazelcast Node.js Client to secure the connection between the client and the members. Near Cache Support for C++ Client Configurable near cache support for IMap in the Hazelcast C++ client. AWS Discovery for C++ Client AWS cluster discovery implementation for Hazelcast C++ Client.
  • 77.
    © 2017 HazelcastInc. Confidential & Proprietary 77 New Hazelcast IMDG 3.8 Cloud & Plugins Features Features Description Pivotal® CloudFoundry Integration Hazelcast IMDG on Pivotal CloudFoundry as a Tile. IBM DynaCache API Support A WebSphere DynamicCache provider, which could be used either by traditional WebSphere or WebSphere Liberty Profile applications. Discovery SPI based AWS Module Discovery SPI compatible hazelcast-aws module. AWS users would take advantage of ZONE_AWARE (availability zone aware partition groups) without using Apache jclouds. Striim Hot Cache Adapter HotCache functionality with Striim based on Change Data Capture ("CDC") of underlying database updates. Hibernate 5.2 Support Improvements in distributed L2 cache for Hibernate to support Hibernate 5.2.
  • 78.
    © 2017 HazelcastInc. Confidential & Proprietary 78 Hazelcast 3.9 Roadmap August 2017
  • 79.
    © 2017 HazelcastInc. Confidential & Proprietary 79 3.9 Features Features Description Near Cache: Key Store-by- Reference Option Key store-by-reference option for IMap and Hazelcast JCache Near Cache. Dynamic Creation of Specificly Configured Distributed Objects Ability to define new data structures with custom configuration on the fly. Gigantic Cache Migration Enhancements Improvements in partition migration process to allow better gigantic cache migrations. Lazy Initiation for Java Client An async client factory to connect lazily and in a non-blocking manner. Client Support for User Code Deployment Dynamic Classloading from clients for User Code Deployment. Query Async API Async query methods on the IMap
  • 80.
    © 2017 HazelcastInc. Confidential & Proprietary 80 3.9 – Security Features Features Description Mutual TLS Auth Java Client Allow a member to authenticate a client using a certificate installed in the client Mutual TLS Auth C# Client Allow a member to authenticate a client using a certificate installed in the client Management Center TLS TLS from members to Man Center. TLS from browser. TLS for REST endpoint TLS for REST endpoint on Members General Security Hardening Implement things like password blacklisting, password complexity requirements, toggling off of non-secure mechanisms when secure version is enabled and so on Fast SSL/TLS Make SSL/TLS almost as fast as unencrypted.
  • 81.
    © 2017 HazelcastInc. Confidential & Proprietary 81 Fast SSL/TLS
  • 82.
    © 2017 HazelcastInc. Confidential & Proprietary 82 New Hazelcast IMDG 3.9 Enterprise Features Features Description Dynamic Addition of WAN Endpoints Ability to add WAN Targets to a running cluster. NearCache Stats in Management Center Provision of showing data inside Near-Cache of client or member, similar to IMap/Cache in Management Center. Search Indexes in HD Memory Support for sorted and unsorted indexes in Hazelcast HD Memory Store.
  • 83.
    © 2017 HazelcastInc. Confidential & Proprietary 83 New Hazelcast IMDG 3.9 Clients Features Features Description Simulator Integrations of C++, Python, and Node.js Clients Integration of C++, Python, and Node.js clients into Hazelcast Simulator. SSL Performance Improvements for .NET Client Large performance increase in SSL when using the .NET Client. Projections for .NET Client Add projections (returning attributes rather than full values) to .NET. Paging Predicate for .NET Client Support for paging predicates in Hazelcast .NET Client. ExecutorService for C++ Client Ability to use the distributed ExecutorService from Hazelcast C++ Client. Near Cache for C++ Client Add near cache to C++ client TLS for C++ Client Add TLS encryption for C++ Clients communicating with members
  • 84.
    © 2017 HazelcastInc. Confidential & Proprietary 84 New Hazelcast IMDG 3.9 Cloud & Plugins Features Features Description Google Cloud Platform Integration SPI Plugin for GCP for Cloud Discovery with a support for partition groups. AWS support for C++ Client We have comprehensive cloud support for Java. Adding AWS to C++. Stats for Hibernate LocalRegionCache Ability to provide statistics for Hibernate LocalRegionCache. Web Sessions for WebSphere Liberty Web sessions replication plugin for WebSphere Liberty.
  • 85.
    © 2017 HazelcastInc. Confidential & Proprietary 85 New Hazelcast IMDG 4.0 Features Features Description Java 8 Language Level Lift the language level for Hazelcast members and their API to Java 8. Add a new Java 8 client. Continue to maintain the Java 6 compatible client. This will allow us to take advantage of Java 8 language features. Java 9 Modularisation Java 9 modularisation for Hazelcast IMDG, which introduces Jigsaw - a new modular systems with a strict enforcement of module boundaries. Proper Separation of Public and Private APIs Re-design of the whole private/public API to keep as much private as possible. Public and Stable SPI Make SPI a second public and stable API. This work includes many refactorings, client protocol enhancements, documentation efforts, and reference implementations. EntryProcessor Contract Formalisation Optimization of the EntryProcessor which would bring the IMap behaviour inline with the JCache behaviour. Unification of Client / Member Behaviour in Exceptional States Equalization of exceptions in client-side and member-side so that clients can throw the same exceptions as server whenever possible. This work also includes a documentation effort on exceptional cases for each method.
  • 87.
    What’s Hazelcast Jet? •General purpose distributed data processing framework • Based on Direct Acyclic Graph to model data flow • Built on top of Hazelcast • Comparable to Apache Spark or Apache Flink 8 7
  • 88.
  • 89.
  • 90.
  • 91.
    Service Offerings Hazelcast (ApacheLicensed) • Professional Subscription – 24x7 support* Hazelcast Enterprise Support • Available with Hazelcast Enterprise software subscription - 24x7 support* Additional Services • Development Support Subscription – 8x5 support* • Simulator TCK • Training • Expert Consulting • Development Partner Program * All subscriptions include Management Center
  • 92.
    © 2016 HazelcastInc. Confidential & Proprietary 92 Support Subscriptions What’s Included 100% SUCCESS RATE ON CUSTOMER ISSUES: “As usual, the response was timely beyond expectations, and very good technical content returned. Exemplary support, hard to find in any company…” - Fortune 100 Financial Services Customer ENTERPRISE HD ENTERPRISE PROFESSIONAL OPEN SOURCE SUPPORT WINDOW 24/7 24/7 24/7 RESPONSE TIME FOR CRITIAL ISSUES 1 Hour 1 Hour 2 Hours SUPPORTED SOFTWARE Hazelcast & Hazelcast Enterprise Hazelcast & Hazelcast Enterprise Hazelcast SUPPORT CONTACTS 4 4 2 SUPPORT CHANNELS Email, IM & Phone Email, IM & Phone Email, IM & Phone PATCH LEVEL FIXES REMOTE MEETINGS (via GoToMeeting) CODE REVIEW (with a Senior Solutions Architect) 2 Hours 2 Hours 2 Hours QUARTERLY REVIEW OF FEATURE REQUES* QUARTERLY REVIEW OF HAZELCAST ROADMAP*
  • 93.
    © 2016 HazelcastInc. Confidential & Proprietary 9 3 Best In Class Support Support from the Engineers who wrote the code SLA Driven – 100% attainment of support response time Follow the Sun Portal, Email and Phone access Go Red, Go Green. Reproduction of issues on Simulator. Proof of fix on Simulator. Periodic Technical Reviews Meet your production schedule and corporate compliance requirements Ensure the success of your development team with training and best practices
  • 94.
  • 95.
    © 2016 HazelcastInc. Confidential & Proprietary 95 Support Testimonials
  • 96.
    © 2016 HazelcastInc. Confidential & Proprietary 96 Release Lifecycle • Regular Feature release each 4-5 months, e.g. 3.3, 3.4, 3.5 • Maintenance release approximately each month with bug fixes based on the current feature release, e.g. 3.4.1 • For older versions, patch releases made available to fix issues • Release End of Life per support contract
  • 97.