October 20th, Q&A session: Get you issues solved and questions answered!

GitHub logo
Edit

Thin Clients Overview

Overview

A thin client is a lightweight Ignite client that connects to the cluster via a standard socket connection. It does not become a part of the cluster topology, never holds any data, and is not used as a destination for compute grid calculations. What it does is simply establish a socket connection to a standard Ignite node and perform all operations through that node.

Thin clients are based on the binary client protocol, which makes it possible to support Ignite connectivity from any programming language.

Ignite provides the following thin clients:

Thin Client Features

The following table outlines features supported by each client.

Thin Client Feature Java .NET C++ Python Node.js PHP

Scan Query

yes

yes

No

yes

yes

yes

Scan Query with a filter

yes

yes

No

No

No

No

SqlFieldsQuery

yes

yes

No

yes

yes

yes

Binary Object API

yes

yes

No

No

yes

yes

Async Operations

No

yes

No

yes

yes

yes

SSL/TLS

yes

yes

yes

yes

yes

yes

Authentication

yes

yes

yes

yes

yes

yes

Partition Awareness

yes

yes

yes

yes

yes

No

Failover

yes

No

yes

yes

yes

yes

Transactions

yes

No

No

No

No

No

Cluster API

yes

yes

No

No

No

No

Cluster discovery

No

yes

No

No

No

No

Compute

yes

yes

No

No

No

No

Service invocation

yes

No

No

No

No

No

Client Connection Failover

All thin clients (except for the .NET thin client) support a connection failover mechanism, whereby the client automatically switches to an available node in case of the current node or connection failure. For this mechanism to work, you need to provide a list of node addresses you want to use for failover purposes in the client configuration. Refer to the specific client documentation for more details.

Partition Awareness

As explained in the Data Partitioning section, data in the cluster is distributed between the nodes in a balanced manner for scalability and performance reasons. Each cluster node maintains a subset of the data and the partition distribution map, which is used to determine the node that keeps the primary/backup copy of requested entries.

Partition awareness allows the thin client to send query requests directly to the node that owns the queried data.

Without partition awareness, an application that is connected to the cluster via a thin client executes all queries and operations via a single server node that acts as a proxy for the incoming requests. These operations are then re-routed to the node that stores the data that is being requested. This results in a bottleneck that could prevent the application from scaling linearly.

Without Partition Awareness

Notice how queries must pass through the proxy server node, where they are routed to the correct node.

With partition awareness in place, the thin client can directly route queries and operations to the primary nodes that own the data required for the queries. This eliminates the bottleneck, allowing the application to scale more easily.

With Partition Awareness
Warning

Note that presently you need to provide addresses of all the server nodes in the connection properties. This also means that if a new server node joins the cluster, you should add the server’s address to the connection properties and reconnect the thin client. Otherwise, the thin client will not be able to send direct requests to this server. This limitation is planned to be addressed before the GA release of the feature.

Partition Awareness is available for the Java, .NET, C++, Python, and Node.js thin clients. Refer to the documentation of the specific client for more information.

Authentication

All thin clients support authentication in the cluster side. Authentication is configured in the cluster configuration, and the client simply provide user credentials. Refer to the documentation of the specific client for more information.

Cluster Configuration

Thin client connection parameters are controlled by the client connector configuration. By default, Ignite accepts client connections on port 10800. You can change the port, connection buffer size and timeout, enable SSL/TLS, etc.

Configuring Thin Client Connector

The following example shows how to configure thin client connection parameters:

<bean class="org.apache.ignite.configuration.IgniteConfiguration" id="ignite.cfg">
    <property name="clientConnectorConfiguration">
        <bean class="org.apache.ignite.configuration.ClientConnectorConfiguration">
            <property name="port" value="10000"/>
        </bean>
    </property>
</bean>
ClientConnectorConfiguration clientConnectorCfg = new ClientConnectorConfiguration();
// Set a port range from 10000 to 10005
clientConnectorCfg.setPort(10000);
clientConnectorCfg.setPortRange(5);

IgniteConfiguration cfg = new IgniteConfiguration().setClientConnectorConfiguration(clientConnectorCfg);

// Start a node
Ignite ignite = Ignition.start(cfg);
var cfg = new IgniteConfiguration
{
    ClientConnectorConfiguration = new ClientConnectorConfiguration
    {
        // Set a port range from 10000 to 10005
        Port = 10000,
        PortRange = 5
    }
};

var ignite = Ignition.Start(cfg);
This API is not presently available for C++. You can use XML configuration.

The following table describes some parameters that you may want to change.

Parameter Description Default Value

thinClientEnabled

Enables or disables thin client connectivity.

true

port

The port for thin client connections.

10800

portRange

This parameters sets a range of ports for thin client connections. For example, if portRange = 10, thin clients can connect to any port from range 10800–18010. The node tries to bind to each port from the range starting from the port until it finds an available one. If all ports are unavailable, the node won’t start.

100

sslEnabled

Set this property to true to enable SSL for thin client connections.

false

See the complete list of parameters in the ClientConnectorConfiguration javadoc.

Enabling SSL/TLS for Thin Clients

Refer to the SSL for Thin Clients and JDBC/ODBC section.