IDC’s report, “Choosing the Right Database Technology in the Age of Digital Transformation” highlights the expanse of data management options as well as how that can cause confusion. Having clarity when looking to build new data applications or modernize workloads already in existence is essential. Three of the tips they provide to help illuminate the right solution for your needs are to select solutions that have:
- Roadmaps which fit your future plans
- Partners that help round out the overall solution
- Technologies with a large talent pool internally and externally
Of course, this is more easily said than done. IDC is quick to dismiss jack-of-all-trade types of solutions in favor of specialized solutions across edge and IoT, data science, streaming, advanced analytics, data lakes, data warehouses, and transactional processing. To help, we’ll explore how to follow IDC’s advice while still selecting technologies suited to a business’s individual needs.
Roadmaps and the future
Dynamic and shifting markets can make predicting future technology needs spotty if not impossible. This makes finding solutions with a good roadmap or path forward even more essential. Make sure that the solution being chosen is structured to grow as new technologies are created and has a clear path forward on sure bets like Artificial Intelligence (AI).
A good example of this can be found in IBM Cloud Pak for Data. Many of IBM’s specialized data management solutions are available through Cloud Pak for Data, including a database, data warehouse, and fast data option. A business can add additional capabilities easily without changing the underlying structure. Moreover, technologies across the Journey to AI can be added such as IBM Watson OpenScale. Whether data management related or otherwise, all can be added by purchasing additional virtual processor cores (VPCs). It is also hardware and cloud agnostic – built on Red Hat OpenShift Container Platform, it can run wherever Red Hat and Linux containers are supported. If future demands necessitate it, selecting new locations will be much simpler.
One shift that can be counted on is the rise of AI. Data management solutions should be both built for AI, making the development of AI applications more effective and efficient, as well as powered by AI, using AI to improve the data management process itself. Data virtualization and support for popular data science languages are two components of this which will be discussed later. However, features like natural language querying to explore data more easily and confidence-based querying to provide probability of accuracy alongside responses have the power to help improve insights in the upcoming years. Data management solutions should have a plan to incorporate these features.
The partner ecosystem is critical to expanding the functionality of data management solutions. Because companies may specialize in different areas of data management, having strong partner relationships opens access to additional functionality in a tightly integrated way.
For example, companies that partner with Cloudera can offer robust data lake implementations alongside what their own database, data warehouse, or other specialties. Similarly, open source specialists like MongoDB can lend their own expertise where NoSQL is required.
However, these companies should not rely fully on their partners; they must provide additive value on top of what the partner delivers. Where Cloudera is concerned, solutions like IBM Db2 Big SQL add value through improved SQL connectivity. Open source solutions can be bolstered by expert support that can speak to both the primary solution, the partner solution, and how both integrate.
Unfortunately, all of the preferred data management options may not be partnered with one another. In these cases, data virtualization, which provides a single view of and access to all data sources without an ETL (extract, transform, load) process, is vital. Businesses can choose the specialized solutions they want and continue to have access to it no matter where the data happens to live.
Internal and external talent
The best way for data management to ensure that there is enough talent available to leverage the technology is to meet the users on their own terms. In other words, they should allow them to use the languages, data formats, and libraries that they are familiar with.
For data scientists and application developers, those languages include Python, JSON, GO, Ruby, PHP, Java, Node.js, Sequelize, and Jupyter Notebooks. Offering these options helps existing employees continue using what they already know, opens up a larger pool of external talent which can be hired, and makes it more likely that code examples made by peers will be available.
SQL is another common language that should be leveraged to make the integration of newer technologies like Graph and blockchain easier. Previously, data would need to be extracted out of relational systems and put into a graph database prior to analysis. However, it is now possible to run graph applications from relational databases and query graph data directly using SQL. This helps eliminate wasted time and potentially reduce costs that would have been incurred with ETL.
With blockchain, data that was previously hard to access by design can now be used with SQL without altering the blockchain itself. Therefore, it can easily be used alongside other data sources to deliver a more complete set of insights. Consider combining blockchain data with weather data. Doing so may help companies transporting goods more accurately determine why there may have been delays on route. And, as with Graph, ETL is no longer necessary in order to save time and costs.
Discover additional tips by reading IDC’s report “Choosing the Right Database Technology in the Age of Digital Transformation” or learn more about a forward-looking database by reading the eBook, Db2: The AI Database.
If you have questions about these tips or other hybrid data management topics, our experts would also be happy to have a free, 30-minute discussion with you one-on-one.