Tips for building an advanced data platform #data #building : #6/10

Tips for building an advanced data platform #data #building : #6/10

Tip #6: Consider when to use a SaaS, an open-source solution or build your own

There is no tool that fits all use-cases. Whether it’s a data warehouse solution, ETL or visualization tool, the choice of which one to use will depend on the goals, needs and current situation in your organization and data department.

Save your team and company from future troubles by thoroughly evaluating your situation and available tools in advance.

  • List your use-cases and requirements
  • Summarize advantages and disadvantages of each tool
  • Evaluate compatibility with your data and tech stack, consider usability
  • Narrow the list down to possible matches
  • Subscribe to the trial period of a SaaS service or install an open-source solution on a test machine, and thoroughly evaluate each tool on use-cases that are as close to real data volume and velocity as possible

Most likely there will be a need to make a trade-off — consider the one that has the least impact. For example, existing ETL tools such as Fivetran or Airbyte might easily handle many various simple batch data sync workflows, but they might not be able to process huge amounts of data in short time fast enough. Research if there is a solution that can handle both use-cases, evaluate if it makes sense to add a custom solution, or go with a completely different option.

This is the part where it’s easy to go wrong and then suffer from the consequences of that choice when the tool already has too many dependencies in your stack. A well thought out choice from the start can save you from a costly and lengthy migration later.

要查看或添加评论,请登录

Dr. RVS Praveen Ph.D的更多文章

社区洞察

其他会员也浏览了