Announcement

Collapse
No announcement yet.

Article on modernization of FTDNA

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Article on modernization of FTDNA

    I saw this posted today in the ISOGG newsgroup, by Debbie Kennett (blogger and ISOGG member):
    "FTDNA will be modernising their family trees and genome database using graph technology from DataStax:

    https://www.businesswire.com/news/ho...ered-DataStax/"
    If anyone out there is more well-versed in IT talk than I am, could you please read and give us all an assessment of what the implications are for FTDNA?

  • #2
    "Modernization"? Graphs have been fashionable, both on and off the internet, for a long time. Perhaps it is an "upgrade", or even an "improvement". Whether new graphs are useful or informative, or just so much window dressing, remains to be seen. Depending on how they are done, too, they could impede usability by taking too long to load, or not displaying properly on some browsers.

    For many years, I noticed that essentially all French doctoral dissertations in the social sciences included at least one bubble chart, usually showing something like historical prices of grain on a map of France. These charts were rarely germane to the argument, but they were evidently de rigueur, because everybody had one!

    It all depends: a graph of what? Time will tell!

    Comment


    • #3
      I see the link doesn't work in my post above, so let's try again:
      https://www.businesswire.com/news/home/20180717005144/en/Gene-Gene%E2%80%99s
      -Modern-Customer-Experience-Powered-DataStax
      /

      John, I took a look at the article again, and see that part of DataStax's job for FTDNA will be to "modernize its database infrastructure and offer a consistent and reliable platform for its users. With family trees able to grow to hundreds of thousands of people, Gene by Gene needed search and compute capabilities that could scale quickly to support a seamless customer experience."

      They will be using Apache Cassandra, which is an alternative to SQL. I only understand these types of things in general terms, as I am not a programmer. But it sounds like the changes/upgrades DataStax will be doing will help with speed and access for FTDNA's databases, including the family trees (and family trees are a type of chart). I hope these changes will also allow new features for analyzing our DNA results, but that is not mentioned in the article.

      Edited to add one person's opinion: I just saw a reply to Debbie Kennett's post, by another person on the ISOGG mailing list, which mentions a different quote from the DataStax article/press release: that FTDNA was already using Apache Cassandra to "future-proof its technology." The poster (who seems to know more about IT than I do) indicates that Cassandra looks to be very well-suited to the demands of FTDNA for data. He regards this updating as good news, and that besides fixing any family tree problems, it may get rid of other FTDNA website glitches. He does warn that software changes don't always go smoothly, so there may be some bumps as it is introduced.

      I hope his assessment is correct! Fingers crossed.
      Last edited by KATM; 19th July 2018, 08:03 AM.

      Comment


      • #4
        Graph Databases

        Here is a Wikipedia article on graph databases that is reasonably understandable for folks with a working understanding of other types of databases, such as relational.

        https://en.wikipedia.org/wiki/Graph_database

        Dave Bechberger, Chief Software Architect at Gene by Gene since April 2017, has a YouYube presentation entitled "What are Graph Databases and Why Do I Care" at https://youtu.be/qiqLhlG4CkU that you might find useful - it's about an hour and I have not viewed it yet. Dave has other websites and links (Google him) that lead to more information.

        An even higher level explanation of graph databases is that it is a different database structure. That is "more better" for data like pedigree charts and DNA matches.

        It is well-suited to databases, like genealogy databases, where bits of data (names, life facts, locations, etc) can have different relationships to each other. This database structure should allow FTDNA customers to see faster access to data, should allow FTDNA to update their database more quickly, and allow FTDNA to grow in size more easily. It should also allow them to offer us new ways to explore and visualize the data - both pedigree charts and DNA results information.

        The genetic genealogy business and its forensic implications is really heating up. For example, Living DNA is promising to roll out pedigree charts this year based on DNA test results with very minimal genealogical information required. For FTDNA to stay relevant in the face of these technological advances, they have to "pave the road" and this database conversion is a logical step.


        I have managed IT projects of this scope and scale - it is a huge undertaking. I look forward to a progress report at the fall symposium.
        Last edited by liverman; 19th July 2018, 08:20 AM.

        Comment


        • #5
          Agreed. Migration to a new database structure (and new to the industry) can be a risky undertaking. I've seen many such projects, many failed, some succeeded. Those that failed often had management who did not respond to the "what about X?" questions from the subject matter experts -- and, eventually, predictably, X came back to bite them, another million $$ or so wasted.

          Comment


          • #6
            Originally posted by liverman View Post
            The genetic genealogy business and its forensic implications is really heating up. For example, Living DNA is promising to roll out pedigree charts this year based on DNA test results with very minimal genealogical information required. For FTDNA to stay relevant in the face of these technological advances, they have to "pave the road" and this database conversion is a logical step.
            I agree & it is good especially since FTDNA advertises DNA testing mainly for Genealogy purposes...so having the ability for detailed and large family trees only makes sense.

            Comment


            • #7
              Likely moving the conference from November 2018, to March 2019 is related to the upgrade project.


              Mr. W.

              Comment

              Working...
              X