This week began with a compliment from Google that made me so proud on behalf of Vertica! On Wednesday, July 25, at the Google Cloud Next 2018 event, Google launched a Machine Learning (ML) beta with two algorithms – linear and logistic regression – and also confirmed that these new machine learning functions could be accessed using standard SQL via Google BigQuery, opening the door for database administrators and business analysts who are very fluent in SQL to capitalize on in-database machine learning. Google explained that specialized machine learning platforms require exporting data from an organization’s data warehouse, and this often requires down sampling, reformatting, and significant delays, making predictive analytics more likely to accept compromises on accuracy and actionable results. Google also announced some new geospatial capabilities in “public alpha” (not really sure what that means!).
It is a proud moment when Vertica’s industry leadership in predictive analytics powered by in-database machine learning and advanced analytical functions including geospatial, time series, pattern matching, projections, and more are copied by one or more of our less mature competitors. As we all know, Vertica has had in-database machine learning, including a much larger selection of algorithms, as well as geospatial analytics and more for years. And need I mention that none of these capabilities in Vertica are either alpha or beta? They are generally available and production ready! I was also very honored to see that Google now understands what Vertica has been advocating for years – integrating machine learning functions where the data resides allows for more accurate and efficient model training and evaluation that then produce more accurate and actionable results. I expect many of our other less mature competitors in the market to see the light as well and follow Vertica’s lead.
Of course, cloud providers like Amazon, Azure, and Google all face a common challenge. Whatever “Vertica-like” functions they add, whatever “Vertica-like” capabilities they integrate and whatever Vertica strategy and value that they endorse, they can and will only do so for data on their own cloud platforms. Vertica, as we all know, is not limited to a single underlying infrastructure or to a single location for all the data. Vertica runs on AWS, Azure, and Google and can apply all of its in-database machine learning and advanced analytics to data on all three of the major cloud platforms. Vertica can also analyze data stored in Hadoop and AWS S3 without the need to move that data. Vertica can read and write ORC and Parquet file formats regardless of the storage platform on which they reside. This is an element of Vertica’s market leadership that cannot be duplicated but is definitely envied by many of our competitors … and highly valued by our global customers!
Every industry is facing the reality that historical reporting is a thing of the past. Companies who rely only on history to make decisions will find themselves in history books but not on the NYSE or FTSE or S&P 500. As Ray Wang from Constellation Research has said, “Digital Darwinism is unkind to those who wait” – one of my very favorite quotes. Digital Darwinism has not only moved past historical reporting and traditional business intelligence but it is rapidly moving to a new set of expectations for predictive analytics as well – accuracy by analyzing ALL the data and action by analyzing in REAL TIME at scale. This is a huge win for the business analyst community who have been frustrated by the system limitations that hamper their impact on their business responsibilities. But not anymore – Vertica to the rescue!