Hello @gang.chen1

I have translated the question using Goole translate, but I am not sure if it is correct.

Question 1: What is the function of the CompressionType configuration item in Channel? What is the recommended configuration?

How to configure the DirectWrite configuration item in Channel?

What is the recommended configuration for the TcpNodelay configuration item in the Channel?

Answer:

The CompressionType configuration is for setting a message compression between the API and server. The supported values are as follows:

• None (Default)
• ZLib
• LZ4

The compression type and compression level must be supported on the server side too. I highly recommend the client contact their Market Data team to verify their RTDS server configurations.

About the "DirectWrite" and "TcpNodelay" recommendations, there is no golden number suggestion. I recommend the client test each setting value to find the one that matches their requirements.

Question 2: From the documentation and examples, the <Host> tag is not configured with multiple addresses, and the <Host> tag and <Port> are used together, then if the IPs of two ADSs are configured, this is the way" <ChannelSet value="Channel_1, Channel_2"/> ", or this way " <Host value="localhost1, localhost2"/> "

Answer: The client can set multiple hosts/IPs with the ChannelSet configuration.

• Host 1 in Channel_1
• Host 2 in Channel_2
• Then set Consumer's ChannelSet to Channel_1 and Channel_2

Example:

<ConsumerGroup>
    <ConsumerList>
        <Consumer>
            <Name value="Consumer_2"/>
            <!-- ChannelSet specifies an ordered list of Channels to which OmmConsumer will attempt to -->
            <!-- connect, one at a time, if the previous one fails to connect -->
            <ChannelSet value="Channel_1, Channel_2"/>
            <Dictionary value="Dictionary_2"/>
            <XmlTraceToStdout value="0"/>
        </Consumer>
    </ConsumerList>
</ConsumerGroup>
...
<ChannelGroup>
    <ChannelList>
        <Channel>
            <Name value="Channel_1"/>        
            <ChannelType value="ChannelType::RSSL_SOCKET"/>
            <CompressionType value="CompressionType::None"/>
            <GuaranteedOutputBuffers value="5000"/>
            <TcpNodelay value="1"/>
            <Host value="localhost1"/>
            <Port value="14002"/>
        </Channel>
        <Channel>
            <Name value="Channel_2"/>
            <ChannelType value="ChannelType::RSSL_SOCKET"/>
            <CompressionType value="CompressionType::None"/>
            <GuaranteedOutputBuffers value="5000"/>
            <TcpNodelay value="1"/>
            <Host value="localhost2"/>
            <Port value="14002"/>
        </Channel>
    </ChannelList>
</ChannelGroup>

Hello @gang.chen1

Before I continue, I highly recommend the client check the following resources that might help them understand the Refintiv Real-Time and RTSDK betters:

1. 10 important things you need to know before you write a Refinitiv Real-Time application article that gives all the basic detail about the Refinitiv Real-Time concept for the client.
2. EMA API step-by-step tutorials (Java, C/C++)

Communication process:

Question 1. Market Price Request and Batch Request / Snapshot Request in Domain Model are all market requests. Is there any relationship between them? Can each request be sent/received individually?

Answer: The normal request message, batch request, and snapshot request message are types of the request message that the API sends to the server. There is no relationship between them and the API can send-receive each request individually.

Question 2: According to the characteristics of the market request, can it be understood as follows:

• Streaming Requests: One request, one RIC subscription, continuous push
• Snapshot Request: One request, one RIC subscription, one feedback
• Batch Request: One request, subscribe to a batch of RICs, and push continuously
• Views: One request, subscribe to certain fields in a RIC, and push continuously

Answer: The client's understanding is correct. Please note that the Batch and View requests can be used together too.

Example:

consumer  = EmaFactory.createOmmConsumer(EmaFactory.createOmmConsumerConfig().host("localhost:14002").username("user"));


//Batch
OmmArray arrayBatch = EmaFactory.createOmmArray();
arrayBatch.add(EmaFactory.createOmmArrayEntry().ascii("FB.O"));
arrayBatch.add(EmaFactory.createOmmArrayEntry().ascii("CSCO.O"));


//View
OmmArray arrayView = EmaFactory.createOmmArray();
arrayView.fixedWidth(2);
arrayView.add(EmaFactory.createOmmArrayEntry().intValue(22)); //BID
arrayView.add(EmaFactory.createOmmArrayEntry().intValue(25)); //ASK


//ElementList
ElementList batchView = EmaFactory.createElementList();
batchView.add(EmaFactory.createElementEntry().array(EmaRdm.ENAME_BATCH_ITEM_LIST, arrayBatch));
batchView.add(EmaFactory.createElementEntry().uintValue(EmaRdm.ENAME_VIEW_TYPE, 1));
batchView.add(EmaFactory.createElementEntry().array(EmaRdm.ENAME_VIEW_DATA, arrayView));
            
consumer.registerClient(EmaFactory.createReqMsg().serviceName("ELEKTRON_DD").payload(batchView), appClient);

Question 3: When consumers send messages to the outside world, they generally use three message types: ReqMsg / PostMsg / GenericMsg. What is the difference between these three messages?

• ReqMsg: Request Message. The application sends this message to the server to request something.
• PostMsg: Post Message. The application sends this message to the server to push content upstream. The upstream can be that RTDS Server, ATS, or even Refinitiv Real-Time (through RCC).
• GenericMsg: Generic Message, a bi-directional message without any implicit interaction semantics associated with it, hence the name generic. It is used by some customized data domains for a specific purpose.

Hello @gang.chen1

Question 4: What is the relationship between Private Streams and Streaming Requests/Snapshot Requests? Can the made request be set to be a private stream?

Answer: A private stream is a stream that contains data this is not to be aggregated, stored, or recovered by the infrastructure (i.e. RTDS). Data exchanged on private streams are exclusively between only two endpoints; Provider and Consumer and have not been shared with other Consumers or Providers.

This virtual private connection can be either a direct connection, through the RTDS, or via a cascaded set of platforms.

All types of requests, functionality, and Domain Models can flow across a private stream including (but not limited to):

• Streaming Requests
• Snapshot Requests
• Posting
• Generic Messages
• Batch Requests
• Views
• All Refinitiv Domain Models & Custom Domain Models

Question 5: Is the default value of privateStream false?

Answer: The request message is not a private stream by default. The application can set the ReqMsg.privateStream() to enable a private stream.

Please note that the Provider application needs to support a private stream too.

Example from EMA ex220_MBP_PrivateStream example.

consumer.registerClient(EmaFactory.createReqMsg().domainType(EmaRdm.MMT_MARKET_BY_PRICE)
    .serviceName("DIRECT_FEED").name("BBH.ITC")
    .privateStream(true), appClient);

Question 6: What is the relationship between Private Streams / Tunnel Streams? In what scenarios do I need to enable Private Streams / Tunnel Streams?

Answer: A tunnel stream is a private stream with additional behaviors, such as end-to-end line of sight for authentication and guaranteed delivery.

The normal streaming consumer applications with the deployed RTDS connection scenario do not use Private Stream and Tunnel Stream.

• In general, the use case of the private stream is to allow the provider and consumer to target content to one another directly. The content can be only specific for only a user/client; for example, only data that a user has permission to access.
• The Tunnel Stream is currently used when connecting the consumer application to the Refinitiv Real-Time over secure internet such as Refinitiv Real-Time -- Optimized (RTO) or Refinitiv Real-Time Contribution (RCC).

Hope this helps.