Skip to Content

Hadoop Integration with SAP Leonardo Cloud Platform

Hi Experts,

We have a Hadoop Database and a cloud connector on our landscape. We want to connect the database to Leonardo Cloud Platform Cockpit. We maintained the hadoop internal and virtual host and port in the cloud connector and tried reaching the cloud from hadoop, but were unable to integrate. Can someone please suggest options available to us if we want to integrate Hadoop with SAP Leonardo Cloud Platform. Could CPI/PI be used as a middleware ? Could SAP Vora connect to Leonardo ?

Add comment
10|10000 characters needed characters exceeded

  • Follow
  • Get RSS Feed

4 Answers

  • Mar 28 at 07:11 PM

    From your description it sounds like SAP Data Hub might be a potential option.

    It does support HDFS as one of the sources: https://help.sap.com/viewer/29ff74dc606c41acad117003f6034ac7/latest/en-US/d77575d3b41f41ef8eee5ade46ff7d84.html

    It can use Leonardo ML services as built-in operators: https://help.sap.com/viewer/29ff74dc606c41acad117003f6034ac7/2.4.latest/en-US/e6e40bbf46ae4acdbae4b09aee98e6af.html

    You would just need to dig if this is suitable in your environment.

    Add comment
    10|10000 characters needed characters exceeded

  • Mar 26 at 01:57 PM

    Hi,

    If you have SDI/ Data Hub in your landscape it is very easy to connect Hadoop data base and fetch records.

    Add comment
    10|10000 characters needed characters exceeded

  • Mar 26 at 04:56 PM

    Hello Anil, Thanks for the input !!

    Can we connect to SAP cloud servers from data hub installed in our landscape ? If you could you please share some documentation or links regarding same, it would be very helpful for us.

    Hello Witalij, Thanks for the reply !!

    We have SAP Leonardo Machine learning space in the Cloud Foundry Environment of the "SAP Cloud Platform Cockpit". Our API's are deployed in this space as shown below. We want to connect to this space from Hadoop. Please suggest as to how we need to approach the integration for this scenario.

    //Thanks

    Tapas Tarun


    leo.png (89.5 kB)
    Add comment
    10|10000 characters needed characters exceeded

    • Hi Tapas. I’m still trying to understand the details. ML services is a set of API to call. Do you want to call them from Hadoop’s MapReduce? Or do you want call these APIs for data stored in Hadoop HDFS?

  • Mar 28 at 09:53 AM

    Hello Witalij,

    We are trying to call the API's from the data stored in Hadoop HDFS. We also tried to connect via python using "pyhs2.connect" method from "pyhs2" library. But are unable to make a connection since Hadoop database is in our VPN Landscape.

    Since we have used cloud connector to connect to our On-Prem SAP systems (which also resides on same VPN) from cloud, we maintained the hadoop database's internal host in cloud connecter and tried calling from python, but were not successful. Would there be any other approach to achieve this integration based on your experience.

    //Thanks

    Tapas Tarun

    Add comment
    10|10000 characters needed characters exceeded