Skip to content

Latest commit

 

History

History
1284 lines (878 loc) · 53.9 KB

README.md

File metadata and controls

1284 lines (878 loc) · 53.9 KB

Lightstreamer Kafka Connector

Extend Kafka topics to the web effortlessly. Stream real-time data to mobile and web apps, anywhere. Scale Kafka to millions of clients.

Introduction

Are you struggling to stream Apache Kafka® events to remote mobile and web applications? Lightstreamer Kafka Connector is an intelligent proxy that bridges the gap between Kafka and the web/mobile frontier, tackling the "last mile" challenge and ensuring your data transcends edge and boundary constraints.

Architecture

Features

  • Intelligent streaming & adaptive throttling: Optimize data flow with intelligent bandwidth management. Adapt real-time data streaming over Kafka topics to accommodate the network connection quality of each remote client.

  • Massive scalability: Connect millions of clients without compromising performance. Fanout real-time messages published on Kafka topics efficiently, preventing overload on the Kafka brokers. Check out the load tests performed on Lightstreamer Kafka Connector vs. plain Kafka.

  • Firewall and proxy traversal: Ensure the delivery of real-time data to clients anywhere, navigating through network obstacles. Stream data seamlessly, even through the strictest corporate firewalls.

  • Maximum security: Implement state-of-the-art encryption, authentication, and fine-grained authorization for robust protection.

  • Extensive broker compatibility: Compatible with all leading brokers, including Apache Kafka, Confluent Platform, Confluent Cloud, Redpanda, Aiven and Amazon MSK.

  • Comprehensive client SDK support:

    Client APIs

To address the unique challenges faced by IT and development teams in integrating Kafka streams with web and mobile applications, Lightstreamer Kafka Connector provides a set of advanced functionalities. It is designed to facilitate real-time data distribution and enhance operational efficiency, allowing developers to focus more on core product features rather than on resolving connectivity and scalability concerns. With its straightforward setup and compatibility with existing Kafka infrastructures, the connector simplifies the extension of Kafka capabilities to a broader range of devices and platforms, creating a seamless link between complex backend systems and user-facing applications deployed over the internet.

Quick Start: Set up in 5 minutes

To efficiently showcase the functionalities of Lightstreamer Kafka Connector, we have prepared an accessible quickstart application located in the examples/quickstart directory. This streamlined application facilitates real-time streaming of data from a Kafka topic directly to a web interface. It leverages a modified version of the Stock List Demo, specifically adapted to demonstrate Kafka integration. This setup is designed for rapid comprehension, enabling you to swiftly grasp and observe the connector's performance in a real-world scenario.

Quickstart Diagram

The diagram above illustrates how in this setup, a stream of simulated market events is channeled from Kafka to the web client via Lightstreamer Kafka Connector.

To provide a complete stack, the app is based on Docker Compose. The Docker Compose file comprises the following services:

  1. broker: a Kafka broker, based on the Docker Image for Apache Kafka. Please notice that other versions of this quickstart are availbale in the in the examples directory, specifically targeted to other brokers, including Confluent Cloud, Redpanda Serverless, Redpanda Self-hosted, Aiven, and more.
  2. kafka-connector: Lightstreamer Server with Kafka Connector, based on the Lightstreamer Kafka Connector Docker image example, which also includes a web client mounted on /lightstreamer/pages/QuickStart
  3. producer: a native Kafka Producer, based on the provided Dockerfile file from the quickstart-producer producer sample client

Run

  1. Make sure you have Docker, Docker Compose, and Java 17 (or later) installed on your local machine.

  2. From the examples/quickstart folder, run the following:

    $ ./start.sh
    ...
     ⠏ Network quickstart_default  Created
     ✔ Container broker            Started
     ✔ Container producer          Started
     ✔ Container kafka-connector   Started
    ...
    Services started. Now you can point your browser to http://localhost:8080/QuickStart to see real-time data.
    ...
  3. Once all containers are ready, point your browser to http://localhost:8080/QuickStart.

  4. After a few moments, the user interface starts displaying the real-time stock data.

    Demo

  5. To shutdown Docker Compose and clean up all temporary resources:

    $ ./stop.sh

Installation

This section will guide you through the installation of Kafka Connector to get it up and running in a very short time.

Requirements

  • JDK version 17 or later.
  • Lightstreamer Server version 7.4.2 or later (check the LS_HOME/GETTING_STARTED.TXT file for the instructions).
  • A running Kafka broker or Kafka Cluster.

Deploy

Get the deployment package from the latest release page. Alternatively, check out this repository and run the following command from the kafka-connector-project folder:

$ ./gradlew distribuite

which generates the lightstreamer-kafka-connector-<version>.zip bundle under the kafka-connector-project/deploy folder.

Then, unzip it into the adapters folder of the Lightstreamer Server installation. Finally, check that the Lightstreamer layout looks like the following:

LS_HOME/
...
├── adapters
│   ├── lightstreamer-kafka-connector-<version>
│   │   ├── LICENSE
│   │   ├── README.md
│   │   ├── adapters.xml
│   │   ├── javadoc
│   │   ├── lib
│   │   ├── log4j.properties
│   └── welcome_res
...
├── audit
├── bin
...

Configure

Before starting Kafka Connector, you need to properly configure the LS_HOME/adapters/lightstreamer-kafka-connector-<version>/adapters.xml file. For convenience, the package comes with a predefined configuration (the same used in the Quick Start app), which can be customized in all its aspects as per your requirements. Of course, you may add as many different connection configurations as desired to fit your needs.

To quickly complete the installation and verify the successful integration with Kafka, edit the data_provider block QuickStart in the file as follows:

  • update the bootstrap.servers parameter with the connection string of Kafka:

    <param name="bootstrap.servers">kafka.connection.string</param>
  • optionally customize the LS_HOME/adapters/lightstreamer-kafka-connector-<version>/log4j.properties file (the current settings produce the additional quickstart.log file)

  • configure topic and record mapping:

    since a generic Ligthstreamer client needs to subscribe to one or more items to receive real-time updates, Kafka Connector has to offer proper mechanisms to realize the mapping between Kafka topics and Lightstreamer items.

    The QuickStart factory configuration comes with a simple mapping through the following settings:

    • an item template:

      <param name="item-template.stock">stock-#{index=KEY}</param>

      which defines the general format name of the items a client must subscribe to to receive updates from Kafka Connector. The extraction expression syntax used here - denoted within #{...} - permits the clients to specify filtering values to be compared against the actual contents of a Kafka record, evaluated through Extraction Keys used to extract each part of a record. In this case, the KEY predefined constant extracts the key part of Kafka records.

    • a topic mapping:

      <param name="map.stocks.to">item-template.stock</param>

      which maps the topic stocks to the provided item template.

    This configuration instructs Kafka Connector to analyze every single event published to the topic stocks and check if it matches against any item subscribed by the client as:

    • stock-[index=1]: an item with the parameter index bound to a record key equal to 1
    • stock-[index=2]: an item with the parameter index bound to a record key equal to 2
    • ...

    Kafka Connector will then route the event to all matched items.

    In addition, the following section defines how to map the record to the tabular form of Lightstreamer fields, by using the aforementioned Extraction Keys. In this case, the VALUE predefined constant extracts the value part of Kafka records.

    <param name="field.stock_name">#{VALUE.name}</param>
    <param name="field.last_price">#{VALUE.last_price}</param>
    <param name="field.ask">#{VALUE.ask}</param>
    <param name="field.ask_quantity">#{VALUE.ask_quantity}</param>
    <param name="field.bid">#{VALUE.bid}</param>
    <param name="field.bid_quantity">#{VALUE.bid_quantity}</param>
    <param name="field.pct_change">#{VALUE.pct_change}</param>
    <param name="field.min">#{VALUE.min}</param>
    <param name="field.max">#{VALUE.max}</param>
    <param name="field.ref_price">#{VALUE.ref_price}</param>
    <param name="field.open_price">#{VALUE.open_price}</param>
    <param name="field.item_status">#{VALUE.item_status}</param>

    This way, the routed event is transformed into a flat structure, which can be forwarded to the clients.

You can get more details about all possible settings in the Configuration section.

Connection with Confluent Cloud

If your target Kafka cluster is Confluent Cloud, you also need to properly configure TLS 1.2 encryption and SASL/PLAIN authentication, as follows:

<param name="encryption.enable">true</param>
<param name="encryption.protocol">TLSv1.2</param>
<param name="encryption.hostname.verification.enable">true</param>

<param name="authentication.enable">true</param>
<param name="authentication.mechanism">PLAIN</param>
<param name="authentication.username">API.key</param>
<param name="authentication.password">secret</param>
...

where you have to replace API.key and secret with the API Key and secret generated on the Confluent CLI or from the Confluent Cloud Console.

Connection with Redpanda Cloud

If your target Kafka cluster is Redpanda Cloud, you also need to properly configure TLS 1.2 encryption and SASL/SCRAM authentication, as follows:

<param name="encryption.enable">true</param>
<param name="encryption.protocol">TLSv1.2</param>
<param name="encryption.hostname.verification.enable">true</param>

<param name="authentication.enable">true</param>
<param name="authentication.mechanism">SCRAM-SHA-256</param>
<!-- <param name="authentication.mechanism">SCRAM-SHA-512</param> -->
<param name="authentication.username">username</param>
<param name="authentication.password">password</param>
...

where you have to replace username and password with the credentials generated from the Redpanda Console.

Start

  1. Launch Lightstreamer Server.

    From the LS_HOME/bin/unix-like directory, run the following:

    $ ./start_background.sh
  2. Attach a Lightstreamer Consumer.

    The kafka-connector-utils submodule hosts a simple Lightstreamer Java client that can be used to test the consumption of Kafka events from any Kafka topics.

    Before launching the consumer, you first need to build it from the kafka-connector-project folder with the command:

    $ ./gradlew distribuiteConsumer

    which generates the lightstreamer-kafka-connector-utils-consumer-all-<version>.jar under the deploy-consumer folder.

    Then, launch it with:

    $ java -jar deploy-consumer/lightstreamer-kafka-connector-utils-consumer-all-<version>.jar --address http://localhost:8080 --adapter-set KafkaConnector --data-adapter QuickStart --items stock-[index=1] --fields ask,bid,min,max

    As you can see, you have to specify a few parameters:

    • --address: the Lightstreamer Server address
    • --adapter-set: the name of the requested Adapter Set, which triggers Ligthtreamer to activate Kafka Connector deployed into the adapters folder
    • --data-adapter: the name of the requested Data Adapter, which identifies the selected Kafka connection configuration
    • --items: the list of items to subscribe to
    • --fields: the list of requested fields for the items

Note

While we've provided examples in JavaScript (suitable for web browsers) and Java (geared towards desktop applications), you are encouraged to utilize any of the Lightstreamer client SDKs for developing clients in other environments, including iOS, Android, Python, and more.

  1. Publish Events.

    The examples/quickstart-producer folder hosts a simple Kafka producer to publish simulated market events for the QuickStart app.

    Before launching the producer, you first need to build it. Open a new shell from the folder and execute the command:

    $ cd examples/quickstart-producer
    $ ./gradlew distribuite

    which generates the quickstart-producer-all under the deploy folder.

    Then, launch it with:

    $ java -jar deploy/quickstart-producer-all.jar --bootstrap-servers <kafka.connection.string> --topic stocks

    producer_video

    Publishing with Confluent Cloud

    If your target Kafka cluster is Confluent Cloud, you also need to provide a properties file that includes encryption and authentication settings, as follows:

    security.protocol=SASL_SSL
    sasl.jaas.config=org.apache.kafka.common.security.plain.PlainLoginModule required username="<API.key>" password="<secret>";
    sasl.mechanism=PLAIN
    ...

    where you have to replace <API.key> and <secret> with the API Key and secret generated on the Confluent CLI or from the Confluent Cloud Console.

    $ java -jar deploy/quickstart-producer-all.jar --bootstrap-servers <kafka.connection.string> --topic stocks --config-file <path/to/config/file>

    Publishing with Redpanda Cloud

    If your target Kafka cluster is Redpanda Cloud, you also need to provide a properties file that includes encryption and authentication settings, as follows:

    security.protocol=SASL_SSL
    sasl.jaas.config=org.apache.kafka.common.security.scram.ScramLoginModule required username="username" password="password";
    sasl.mechanism=SCRAM-SHA-256
    #sasl.mechanism=SCRAM-SHA-512
    ...

    where you have to replace username and password with the credentials generated from the Redpanda Console, and specify the configured SASL mechanism (SCRAM-SHA-256 or SCRAM-SHA-512).

    $ java -jar deploy/quickstart-producer-all.jar --bootstrap-servers <kafka.connection.string> --topic stocks --config-file <path/to/config/file>
  2. Check Consumed Events.

    After starting the publisher, you should immediately see the real-time updates flowing from the consumer shell:

    consumer_video

Configuration

As already anticipated, Kafka Connector is a Lightstreamer Adapter Set, which means it is made up of a Metadata Adapter and one or more Data Adapters, whose settings are defined in the LS_HOME/adapters/lightstreamer-kafka-connector-<version>/adapters.xml file.

The following sections will guide you through the configuration details.

Global Settings

adapter_conf['id'] - Kafka Connector identifier

Mandatory. The id attribute of the adapters_conf root tag defines the Kafka Connector identifier, which will be used by the Clients to request this Adapter Set while setting up the connection to a Lightstreamer Server through a LightstreamerClient object.

The factory value is set to KafkaConnector for convenience, but you are free to change it as per your requirements.

Example:

<adapters_conf id="KafkaConnector">

adapter_class

Mandatory. The adapter_class tag, specified inside the metadata_provider block, defines the Java class name of the Metadata Adapter.

The factory value is set to com.lightstreamer.kafka_connector.adapters.KafkaConnectorMetadataAdapter, which implements the internal business of Kafka Connector.

It is possible to provide a custom implementation by extending this class: just package your new class in a jar file and deploy it along with all required dependencies into the LS_HOME/adapters/lightstreamer-kafka-connector-<version>/lib folder.

See the section Customize the Kafka Connector Metadata Class for more details.

Example:

...
<metadata_provider>
    ...
    <adapter_class>your.custom.class</adapter_class>
    ...
</metadata_provider>
...

logging.configuration.path

Mandatory. The path of the reload4j configuration file, relative to the deployment folder (LS_HOME/adapters/lightstreamer-kafka-connector-<version>).

The parameter is specified inside the metadata_provider block.

The factory value points to the predefined file LS_HOME/adapters/lightstreamer-kafka-connector-<version>/log4g.properties.

Example:

...
<metadata_provider>
    ...
    <param name="logging.configuration.path">log4j.properties</param>
    ...
</metadata_provider>
...

Connection Settings

Kafka Connector allows the configuration of separate independent connections to different Kafka brokers/clusters.

Every single connection is configured via the definition of its own Data Adapter through the data_provider block. At least one connection must be provided.

Since Kafka Connector manages the physical connection to Kafka by wrapping an internal Kafka Consumer, several configuration settings in the Data Adapter are identical to those required by the usual Kafka Consumer configuration.

General Parameters

data_provider['name'] - Kafka Connection Name

Optional. The name attribute of the data_provider tag defines Kafka Connection Name, which will be used by the Clients to request real-time data from this specific Kafka connection through a Subscription object.

Furthermore, the name is also used to group all logging messages belonging to the same connection.

Tip

For every Data Adaper connection, add a new logger and its relative file appender to log4j.properties, so that you can log to dedicated files all the interactions pertinent to the connection with the Kafka cluster and the message retrieval operations, along with their routing to the subscribed items. For example, the factory logging configuration provides the logger QuickStart to print every log messages relative to the QuickStart connection:

...
# QuickStart logger
log4j.appender.QuickStartFile=org.apache.log4j.RollingFileAppender
log4j.appender.QuickStartFile.layout=org.apache.log4j.PatternLayout
log4j.appender.QuickStartFile.layout.ConversionPattern=[%d] [%-10c{1}] %-5p %m%n
log4j.appender.QuickStartFile.File=../../logs/quickstart.log

Example:

<data_provider name="BrokerConnection">

Default value: DEFAULT, but only one DEFAULT configuration is permitted.

adapter_class

Mandatory. The adapter_class tag defines the Java class name of the Data Adapter. DO NOT EDIT IT!.

Factory value: com.lightstreamer.kafka_connector.adapters.KafkaConnectorAdapter.

enable

Optional. Enable this connection configuration. Can be one of the following:

  • true
  • false

If disabled, Lightstreamer Server will automatically deny every subscription made to this connection.

Default value: true.

Example:

<param name="enable">false</param>
bootstrap.servers

Mandatory. The Kafka Cluster bootstrap server endpoint expressed as the list of host/port pairs used to establish the initial connection.

The parameter sets the value of the bootstrap.servers key to configure the internal Kafka Consumer.

Example:

<param name="bootstrap.servers">broker:29092,broker:29093</param>
group.id

Optional. The name of the consumer group this connection belongs to.

The parameter sets the value for the group.id key to configure the internal Kafka Consumer.

Default value: Kafka Connector Identifier + Connection Name + Randomly generated suffix.

<param name="group.id">kafka-connector-group</param>

Encryption Parameters

A TCP secure connection to Kafka is configured through parameters with the prefix encryption.

encryption.enable

Optional. Enable encryption of this connection. Can be one of the following:

  • true
  • false

Default value: false.

Example:

<param name="encryption.enable">true</param>
encryption.protocol

Optional. The SSL protocol to be used. Can be one of the following:

  • TLSv1.2
  • TLSv1.3

Default value: TLSv1.3 when running on Java 11 or newer, TLSv1.2 otherwise.

Example:

<param name="encryption.protocol">TLSv1.2</param>
encryption.enabled.protocols

Optional. The list of enabled secure communication protocols.

Default value: TLSv1.2,TLSv1.3 when running on Java 11 or newer, TLSv1.2 otherwise.

Example:

<param name="encryption.enabled.protocols">TLSv1.3</param>
encryption.cipher.suites

Optional. The list of enabled secure cipher suites.

Default value: all the available cipher suites in the running JVM.

Example:

<param name="encryption.cipher.suites">TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA</param>
encryption.hostname.verification.enable

Optional. Enable hostname verification. Can be one of the following:

  • true
  • false

Default value: false.

Example:

<param name="encryption.hostname.verification.enable">true</param>
encryption.truststore.path

Optional. The path of the trust store file, relative to the deployment folder (LS_HOME/adapters/lightstreamer-kafka-connector-<version>).

Example:

<param name="encryption.truststore.path">secrets/kafka-connector.truststore.jks</param>
encryption.truststore.password

Optional. The password of the trust store.

If not set, checking the integrity of the trust store file configured will not be possible.

Example:

<param name="encryption.truststore.password">kafka-connector-truststore-password</param>
encryption.keystore.enable

Optional. Enable a key store. Can be one of the following:

  • true
  • false

A key store is required if the mutual TLS is enabled on Kafka.

If enabled, the following parameters configure the key store settings:

  • encryption.keystore.path
  • encryption.keystore.password
  • encryption.keystore.key.password

Default value: false.

Example:

<param name="encryption.keystore.enable">true</param>
encryption.keystore.path

Mandatory if key store is enabled. The path of the key store file, relative to the deployment folder (LS_HOME/adapters/lightstreamer-kafka-connector-<version>).

Example:

<param name="encryption.keystore.path">secrets/kafka-connector.keystore.jks</param>
encryption.keystore.password

Optional. The password of the key store.

If not set, checking the integrity of the key store file configured will not be possible.

Example:

<param name="encryption.keystore.password">keystore-password</param>
encryption.keystore.key.password

Optional. The password of the private key in the key store file.

Example:

<param name="encryption.keystore.key.password">kafka-connector-private-key-password</param>
Quick Start SSL Example

Check out the adapters.xml file of the Quick Start SSL app, where you can find an example of encryption configuration.

Broker Authentication Parameters

Broker authentication is configured through parameters with the prefix authentication.

authentication.enable

Optional. Enable the authentication of this connection against the Kafka Cluster. Can be one of the following:

  • true
  • false

Default value: false.

Example:

<param name="authentication.enable">true</param>
authentication.mechanism

Mandatory if authentication is enabled. The SASL mechanism type. Kafka Connector accepts the following authentication mechanisms:

  • PLAIN (the default value)
  • SCRAM-SHA-256
  • SCRAM-SHA-512
  • GSSAPI

In the case of PLAIN, SCRAM-SHA-256, and SCRAM-SHA-512 mechanisms, the credentials must be configured through the following mandatory parameters (which are not allowed for GSSAPI):

  • authentication.username: the username
  • authentication.password: the password
PLAIN

Example:

<param name="authentication.enable">true</param>
<param name="authentication.mechanism">PLAIN</param>
<param name="authentication.username">authorized-kafka-user</param>
<param name="authentication.password">authorized-kafka-user-password</param>
SCRAM-SHA-256

Example:

<param name="authentication.enable">true</param>
<param name="authentication.mechanism">SCRAM-SHA-256</param>
<param name="authentication.username">authorized-kafka-usee</param>
<param name="authentication.password">authorized-kafka-user-password</param>
SCRAM-SHA-512

Example:

<param name="authentication.enable">true</param>
<param name="authentication.mechanism">SCRAM-SHA-512</param>
<param name="authentication.username">authorized-kafka-username</param>
<param name="authentication.password">authorized-kafka-username-password</param>
GSSAPI

In the case of GSSAPI authentication mechanism, the following parameters will be part of the authentication configuration:

  • authentication.gssapi.key.tab.enable

    Optional. Enable the use of a keytab. Can be one of the following:

    • true
    • false

    Default value: false.

  • authentication.gssapi.key.tab.path

    Mandatory if keytab is enabled. The path to the kaytab file, relative to the deployment folder (LS_HOME/adapters/lightstreamer-kafka-connector-<version>).

  • authentication.gssapi.store.key.enable

    Optional. Enable storage of the principal key. Can be one of the following:

    • true
    • false

    Default value: false.

    Default value: false.

  • authentication.gssapi.kerberos.service.name

    Mandatory. The name of the Kerberos service.

  • authentication.gssapi.principal

    Mandatory if ticket cache is disabled. The name of the principal to be used.

  • authentication.gssapi.ticket.cache.enable

    Optional. Enable the use of a ticket cache. Can be one of the following:

    • true
    • false

    Default value: false.

Example:

...
<param name="authentication.enable">true</param>
<param name="authentication.mechanism">GSSAPI</param>
<param name="authentication.gssapi.key.tab.enable">true</param>
<param name="authentication.gssapi.key.tab.path">gssapi/kafka-connector.keytab</param>
<param name="authentication.gssapi.store.key.enable">true</param>
<param name="authentication.gssapi.kerberos.service.name">kafka</param>
<param name="authentication.gssapi.principal">[email protected]</param>
...

Example of configuration with the use of a ticket cache:

<param name="authentication.enable">true</param>
<param name="authentication.mechanism">GSSAPI</param>
<param name="authentication.gssapi.kerberos.service.name">kafka</param>
<param name="authentication.gssapi.ticket.cache.enable">true</param>
Quick Start Confluent Cloud Example

Check out the adapters.xml file of the Quick Start Confluent Cloud app, where you can find an example of an authentication configuration that uses SASL/PLAIN.

Quick Start Redpanda Serverless Example

Check out the adapters.xml file of the Quick Start Redpanda Serverless app, where you can find an example of an authentication configuration that uses SASL/SCRAM.

Record Evaluation

Kafka Connector can deserialize Kafka records from the following formats:

  • Apache Avro
  • JSON
  • String
  • Integer
  • Float

and other scalar types (see the complete list).

In particular, Kafka Connector supports message validation for Avro and JSON, which can be specified through:

  • local schema files
  • the Confluent Schema Registry

Kafka Connector allows independent deserialization of keys and values, which means that:

Important

For Avro, schema validation is mandatory, therefore either a local schema file must be provided or the Confluent Schema Registry must be enabled.

record.consume.from

Optional. Specifies where to start consuming events from:

  • LATEST: start consuming events from the end of the topic partition
  • EARLIEST: start consuming events from the beginning of the topic partition

The parameter sets the value of the auto.offset.reset key to configure the internal Kafka Consumer.

Default value: LATEST.

Example:

<param name="record.consume.from">EARLIEST</param>
record.key.evaluator.type and record.value.evaluator.type

Optional. The format to be used to deserialize respectively the key and value of a Kafka record. Can be one of the following:

  • AVRO
  • JSON
  • STRING
  • INTEGER
  • BOOLEAN
  • BYTE_ARRAY
  • BYTE_BUFFER
  • BYTES
  • DOUBLE
  • FLOAT
  • LONG
  • SHORT
  • UUID

Default value: STRING.

Examples:

<param name="record.key.evaluator.type">INTEGER</param>
<param name="record.value.evaluator.type">JSON</param>
record.key.evaluator.schema.path and record.value.evaluator.schema.path

Mandatory if evaluator type is AVRO and the Confluent Schema Registry is disabled. The path of the local schema file relative to the deployment folder (LS_HOME/adapters/lightstreamer-kafka-connector-<version>) for message validation respectively of the Kafka key and the Kafka value.

Examples:

<param name="record.key.evaluator.schema.path">schema/record_key.avsc</param>
<param name="record.value.evaluator.schema.path">schemas/record_value.avsc</param>
record.key.evaluator.schema.registry.enable and record.value.evaluator.schema.registry.enable

Mandatory if evaluator type is AVRO and no local schema paths are specified. Enable the use of the Confluent Schema Registry for validation respectively of the Kafka key and the Kafka value. Can be one of the following:

  • true
  • false

Default value: false.

Examples:

<param name="record.key.evaluator.schema.registry.enable">true</param>
<param name="record.value.evaluator.schema.registry.enable">true</param>
record.extraction.error.strategy

Optional. The error handling strategy to be used if an error occurs while extracting data from incoming deserialized records. Can be one of the following:

  • IGNORE_AND_CONTINUE: ignore the error and continue to process the next record
  • FORCE_UNSUBSCRIPTION: stop processing records and force unsubscription of the items requested by all the clients subscribed to this connection

Default value: IGNORE_AND_CONTINUE.

Example:

<param name="record.extraction.error.strategy">FORCE_UNSUBSCRIPTION</param>

Topic Mapping

Kafka Connector allows the configuration of several routing and mapping strategies, thus enabling the convey of Kafka events streams to a potentially huge amount of devices connected to Lightstreamer with great flexibility.

As anticipated in the Installation section, a Kafka record can be analyzed in all its aspects to extract information that can be:

  • routed to the designated Lightstreamer items
  • remapped to specific Lightstreamer fields
Record Routing (map.<topic>.to)

To configure the routing of Kafka event streams to Lightstreamer items, use at least one parameter map.<topic>.to. The general format is:

<param name="map.<topic-name>.to">item1,item2,itemN,...</param>

which defines the mapping between the source Kafka topic (<topic-name>) and the target items (item1, item2, itemN, etc.).

This configuration enables the implementation of various mapping scenarios, as shown by the following examples:

  • One To One

    <param name="map.sample-topic.to">sample-item</param>

    one-to-one

    This is the most straightforward scenario one may think of: every record published to the Kafka topic sample-topic will simply be routed to the Lightstreamer item sample-item. Therefore, messages will be immediately broadcasted as real-time updates to all clients subscribed to such an item.

  • One To Many

    <param name="map.sample-topic.to">sample-item1,sample-item2,sample-item3</param>

    one-to-many

    Every record published to the Kafka topic sample-topic will be routed to the Lightstreamer items sample-item1, sample-item2, and sample-item3.

    This scenario may activate unicast and multicast messaging, as it is possible to specify which item can be subscribed to by which user or group of users. To do that, it is required to provide a customized extension of the factory Metadata Adapter class (see the example), in which every subscription must be validated against the user identity.

  • Many to One

    <param name="map.sample-topic1.to">sample-item</param>
    <param name="map.sample-topic2.to">sample-item</param>
    <param name="map.sample-topic3.to">sample-item</param>

    many-to-one

    With this scenario, it is possible to broadcast to all clients subscribed to a single item (sample-item) every message published to different topics (sample-topic1, sample-topic2, sample-topic3).

Record Mapping (field.<fieldName>)

To forward real-time updates to the Lightstreamer clients, a Kafka record must be mapped to Lightstreamer fields, which define the schema of any Lightstreamer item.

record-mapping

To configure the mapping, you define the set of all subscribeable fields through parameters with the prefix field.:

<param name="field.fieldName1">extraction_expression1</param>
<param name="field.fieldName2">extraction_expression2<param>
...
<param name="field.fieldNameN">extraction_expressionN<param>
...

The configuration specifies that the field fieldNameX will contain the value extracted from the deserialized Kafka record through the extraction_expressionX. This approach makes it possible to transform a Kafka record of any complexity to the flat structure required by Lightstreamer.

To write an extraction expression, Kafka Connector provides the Data Extraction Language. This language has a pretty minimal syntax, with the following basic rules:

  • expressions must be enclosed within #{...}

  • expressions use Extraction Keys, a set of predefined constants that reference specific parts of the record structure:

    • #{KEY}: the key
    • #{VALUE}: the value
    • #{TOPIC}: the topic
    • #{TIMESTAMP}: the timestamp
    • #{PARTITION}: the partition
    • #{OFFSET}: the offset
  • the dot notation is used to access attributes or fields of record keys and record values serialized in JSON or Avro formats:

    KEY.attribute1Name.attribute2Name...
    VALUE.attribute1Name.attribute2Name...

Important

Currently, it is required that the top-level element of either a record key or record value is:

  • an Object, in the case of JSON format
  • a Record, in the case of Avro format

Such a constraint may be removed in a future version of Kafka Connector.

  • the square notation is used to access:

    • indexed attributes:

      KEY.attribute1Name[i].attribute2Name...
      VALUE.attribute1Name[i].attribute2Name...

      where i is a 0-indexed value.

    • key-based attributes:

      KEY.attribute1Name['keyName'].attribute2Name...
      VALUE.attribute1Name['keyName'].attribute2Name...

      where keyName is a string value.

Tip

For JSON format, accessing a child attribute by dot notation or square bracket notation is equivalent:

VALUE.myProperty.myChild.childProperty
VALUE.myProperty['myChild'].childProperty
  • expressions must evaluate to a scalar value, otherwise an error will be thrown during the extraction process. The error will be handled as per the configured strategy.

The QuickStart factory configuration shows a basic example, where a simple direct mapping has been defined between every attribute of the JSON record value and a Lightstreamer field with the same name. Of course, thanks to the Data Extraction Language, more complex mapping can be employed.

...
<param name="field.timestamp">#{VALUE.timestamp}</param>
<param name="field.time">#{VALUE.time}</param>
<param name="field.stock_name">#{VALUE.name}</param>
<param name="field.last_price">#{VALUE.last_price}</param>
<param name="field.ask">#{VALUE.ask}</param>
<param name="field.ask_quantity">#{VALUE.ask_quantity}</param>
<param name="field.bid">#{VALUE.bid}</param>
<param name="field.bid_quantity">#{VALUE.bid_quantity}</param>
<param name="field.pct_change">#{VALUE.pct_change}</param>
<param name="field.min">#{VALUE.min}</param>
<param name="field.max">#{VALUE.max}</param>
<param name="field.ref_price">#{VALUE.ref_price}</param>
<param name="field.open_price">#{VALUE.open_price}</param>
<param name="field.item_status">#{VALUE.item_status}</param>
..
Filtered Record Routing (item-template.<template-name>)

Besides mapping topics to statically predefined items, Kafka Connector allows you to configure the item templates, which specify the rule needed to decide if a message can be forwarded to the items specified by the clients, thus enabling a filtered routing. The item template leverages the Data Extraction Language to extract data from Kafka records and match them against the parameterized subscribed items.

filtered-routing

To configure an item template, use the parameter item-template.<template-name>:

<param name="item-template.<template-name>"><item-prefix>-<expressions></param>

and then configure the routing by referencing the template through the parameter map.<topic>.to:

<param name="map.<topic>.to">item-template.<template-name></param>

Tip

It is allowed to mix references to simple item names and item templates in the same topic mapping configuration:

<param name="map.sample-topic.to">item-template.template1,item1,item2</param>

The item template is made of:

  • <prefix>: the prefix of the item name

  • <expressions>: a sequence of extraction expressions, which define filtering rules specified as:

    #{paramName1=<extraction_expression_1>,paramName2=<extraction_expression_2>,...}

    where paramNameX is a bind parameter to be specified by the clients and whose actual value will be extracted from the deserialized Kafka record by evaluating the <extraction_expression_X> expression (written using the Data Extraction Language).

To activate the filtered routing, the Lightstreamer clients must subscribe to a parameterized item that specifies a filtering value for every bind parameter defined in the template:

<item-prefix>-[paramName1=filterValue_1,paramName2=filerValue_2,...]

Upon consuming a message, Kafka Connector expands every item template addressed by the record topic by evaluating each extraction expression and binding the extracted value to the associated parameter. The expanded template will result as:

<item-prefix>-[paramName1=extractedValue_1,paramName2=extractedValue_2,...]

Finally, the message will be mapped and routed only in case the subscribed item completely matches the expanded template or, more formally, the following is true:

filterValue_X == extractValue_X for every paramName_X

Example

Consider the following configuration:

<param name=item-template.by-name>user-#{firstName=VALUE.name,lastName=VALUE.surname}</param>
<param name=item-template.by-age>user-#{years=VALUE.age}</param>
<param name="map.user.to">item-template.by-name,item-template.by-age</param>

which specifies how to route records published from the topic user to item templates defined to extract some personal data.

Let's suppose we have three different Lightstreamer clients:

  1. Client A subscribes to:
    • the parameterized item SA1 user-[firstName=James,lastName=Kirk] for receiving real-time updates relative to the user James Kirk
    • the parameterized item SA2 user-[age=45] for receiving real-time updates relative to any 45 year-old user
  2. Client B subscribes to:
    • the parameterized item SB1 user-[firstName=Montgomery,lastName=Scotty] for receiving real-time updates relative to the user Montgomery Scotty
  3. Client C subscribes to the parameterized item SC1 user-[age=37] for receiving real-time updates relative to any 37 year-old user.

Now, let's see how filtered routing works for the following incoming Kafka records published to the topic user:

  • Record 1:

    {
      ...
      "name": "James",
      "surname": "Kirk",
      "age": 37,
      ...
    }
    Template Expansion Matched Subscribed Item Routed to Client
    by-name user-[firstName=James,lastName=Kirk] SA1 Client A
    by-age user-[age=37] SC1 Client C
  • Record 2:

    {
      ...
      "name": "Montgomery",
      "surname": "Scotty",
      "age": 45
      ...
    }
    Template Expansion Matched Subscribed Item Routed to Client
    by-name user-[firstName=Montgomery,lastName=Scotty] SB1 Client B
    by-age user-[age=45] SA2 Client A
  • Record 3:

    {
      ...
      "name": "Nyota",
      "surname": "Uhura",
      "age": 37,
      ...
    }
    Template Expansion Matched Subscribed Item Routed to Client
    by-name user-[firstName=Nyota,lastName=Uhura] None None
    by-age user-[age=37] SC1 Client C

Schema Registry

A Schema Registry is a centralized repository that manages and validates schemas, which define the structure of valid messages.

Kafka Connector supports integration with the Confluent Schema Registry through the configuration of parameters with the prefix schema.registry.

schema.registry.url

Mandatory if the Confluent Schema Registry is enabled. The URL of the Confluent Schema Registry.

Example:

<param name="schema.registry.url">http//localhost:8081</param>

An encrypted connection is enabled by specifying the https protocol (see the next section).

Example:

<param name="schema.registry.url">https://localhost:8084</param>
Encryption Parameters

A secure connection to the Confluent Schema Registry can be configured through parameters with the prefix schema.registry.encryption, each one having the same meaning as the homologous parameters defined in the Encryption Parameters section:

Example:

<!-- Set the Confluent Schema Registry URL. The https protcol enable encryption parameters -->
<param name="schema.registry.url">https//localhost:8084</param>

<!-- Set general encryption settings -->
<param name="schema.registry.encryption.enabled.protocols">TLSv1.3</param>
<param name="schema.registry.encryption.cipher.suites">TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA</param>
<param name="schema.registry.encryption.hostname.verification.enable">true</param>

<!-- If required, configure the trust store to trust the Confluent Schema Registry certificates -->
<param name="schema.registry.encryption.truststore.path">secrets/secrets/kafka.connector.schema.registry.truststore.jks</param></param>

<!-- If mutual TLS is enabled on the Confluent Schema Registry, enable and configure the key store -->
<param name="schema.registry.encryption.keystore.enable">true</param>
<param name="schema.registry.encryption.keystore.path">secrets/kafka-connector.keystore.jks</param>
<param name="schema.registry.encryption.keystore.password">kafka-connector-password</param>
<param name="schema.registry.encryption.keystore.key.password">schemaregistry-private-key-password</param>
Quick Start Schema Registry Example

Check out the adapters.xml file of the Quick Start Schema Registry app, where you can find an example of Schema Registry settings.

Customize the Kafka Connector Metadata Adapter Class

If you have any specific need to customize the Kafka Connector Metadata Adapter class, you can provide your implementation by extending the factory class com.lightstreamer.kafka_connector.adapters.pub.KafkaConnectorMetadataAdapter. The class provides the following hook methods, which you can override to add your custom logic:

  • postInit: invoked after the initialization phase of the Kafka Connector Metadata Adapter has been completed

  • onSubscription: invoked to notify that a user has submitted a subscription

  • onUnsubcription: invoked to notify that a Subscription has been removed

Develop the Extension

To develop your extension, you need the Kafka Connector jar library, which is hosted on Github Packages.

For a Maven project, add the dependency to your pom.xml file:

<dependency>
    <groupId>com.lightstreamer.kafka-connector</groupId>
    <artifactId>kafka-connector</artifactId>
    <version>0.1.0</version>
</dependency>

and follow these instructions to configure the repository and authentication.

For a Gradle project, edit your build.gradle file as follows:

  1. add the dependency:

    dependencies {
        implementation group: 'com.lightstreamer.kafka_connector', name: 'kafka-connector', 'version': '0.1.0'
    }
  2. add the repository and specify your personal access token:

    repositories {
        mavenCentral()
        maven {
            name = "GitHubPackages"
            url = uri("https://maven.pkg.github.com/lightstreamer/lightstreamer-kafka-connector")
                credentials {
                    username = project.findProperty("gpr.user") ?: System.getenv("USERNAME")
                    password = project.findProperty("gpr.key") ?: System.getenv("TOKEN")
                }
        }
    }
    

In the examples/custom-kafka-connector-adapter folder, you can find a sample Gradle project you may use as a starting point to build and deploy your custom extension.

Docs

The docs folder contains the complete Kafka Connector API Specification, already mentioned in the previous section.

Examples

The examples folder contains all the examples mentioned throughout this guide. Furthermore, you may take a look at the Airport Demo, which provides more insights into various usage and configuration options of Kafka Connector.