Connectivity in Cloud Pak for Data 3.5, What, Why, How…
I’m thrilled to announce the recent changes in the new coming version of IBM Cloud Pak for Data, to ease your experience of connecting to a data source.
In previous versions of IBM Cloud Pak for Data, you may have tried to create a connection to a data source, as this is the mandatory first step to even consider getting value out of your data.
For IT teams, it breaks down into 2 challenges-
- Providing quick access to the right data sources for data scientists, data engineers and data analysts
- Being able to define connections to new data sources, without slowing down the response time to LoB or data science users, as outlined in #1 above.
We’ve worked on the user experience, as a whole, to take into account the needs of the various personas of the platform and made several improvements.
Define once and reuse… Platform Connections
In version 3.5, we are introducing the notion of “Platform connections”. Platform connections (Accessible from the Data-> Platform Connections in the navigation menu) will rely only on personal credentials.
An Administrator of the platform will have the ability to define access control for platform connections, based on user credentials. This means, defining the access controls to access the data source, Administrators will have the ability to either assign this role to all the Cloud Pak for Data users, or individually to a subset of users.
All users will have the ability to ‘use’ some platform level connections (in a project, a catalog), but only ‘Editors’ (identified and defined as such in the Access control section by Admins) of Platform connections will be allowed to create new connections.
In other words, if I am an admin and you are a user: you can access a data source I’ve defined (without redefining URL, SSL cert, port….) if you have an account to connect to it
Visible to everyone… A Platform Assets Catalog
All platform connections are stored in the Platforms Assets Catalog, which is available by default and accessible to view by all users of Cloud Pak for Data. It is instantiated as a single tenant instance (not shared across the Cloud Pak for Data instance, in case you have several installed). This can also be used to store other asset types (that you would like to make accessible and visible to anyone). Administrators will have to configure this Platform Assets Catalog, to define who is allowed to contribute to it (definition of the catalog admins, and editors). This catalog is part of the Common Core Services of the platform, and can be used to store any type of assets.
Same connection page everywhere… Unified Experience
We’ve also unified the way users would create connections in the platform (and thus the list of available connectors).
Indeed, in previous versions of the product, the feedback we received was that there were several ways of creating a connection, and it also led to challenges when trying to re-use a connection in a different services. Starting in Cloud Pak for Data 3.5, we’ve defined a common User interface that will be exposed in several areas of the product. For the coming release, you can find this new experience in Notebooks, SPSS Modeler, Watson Knowledge Catalog including Discovery and Data Virtualization. We will continue this alignment in future releases of Cloud Pak for Data.
Last but not least, you will notice when navigating this page, that we’ve enriched the list of connectors by 16 new types of data sources :
Apache Cassandra
Microsoft Azure Blob Storage
Box
Amazon RDS for MySQL
Amazon RDS for PostgreSQL
HTTP
SAP Hana (no driver provided)
IBM Data Virtualization Manager (z/OS)
Microsoft Azure CosmosDB
ElasticSearch
MongoDB
Azure MariaDB
Storage volume
IBM SPSS Analytics Server
Apache Derby
IBM DB2 EventStore
This list adds to the list of connectors documented here https://www.ibm.com/support/producthub/icpdata/docs/content/SSQNUZ_current/cpd/access/data-sources.html
We have also improved the existing set of connectors with additional capabilities. For instance, load support for OData-related connectors (Generic OData and SAP OData), SSL support on some connectors, ability to connect to Amazon S3 GovCloud (using the Amazon S3 connector) or JWT token support, allowing to leverage the platform level authentication mechanism.
In case you couldn’t find an appropriate connector in this list, to connect to your datasource, we are still providing you the option to use your own JDBC drivers, and define your “Generic JDBC connection”, that will rely on those drivers.
I hope you’ll enjoy and see the benefits of this new connectivity experience in Cloud Pak for Data 3.5 !
Want to know more? Watch the Platform Connections demo or join the Cloud Pak Community
