![]() ![]() Postgres and MySQL are both widely used, and why not. Many internal apps or SaaS offerings I’ve seen start on a single database that supports everything, from transactions to analytics. One database for transactions and analytics ![]() As a famous old mentor once said “forewarned, forearmed Grasshopper.” 1. Perhaps if you recognize the stages, and the limitations of Postgres, MySQL, and even many data warehouses in supporting the 1 second SLA, perhaps you can skip a few stages and save yourself some pain. I do hope understanding these stages in advance will help you. There are valid reasons for adopting and then changing the technologies at each stage of growth. I could say just skip some stages, but that’s not how applications evolve. A separate data warehouse for all analytics.One database for transactions and analytics.There have also been a few variants of MySQL, not to mention in-memory caches and other cloud data warehouses.īut almost all of my conversations seem to fit into 3 stages of performance challenges as companies have grown: There are many versions of the Postgres “elephant”, even some analytics-optimized versions like Greenplum, Netezza, ParAccel, and yes, Redshift. There is no one vendor that stands out as being worse than others. The first big elephant in the room during my conversations, the one slowing down analytics, is usually a transactional database. As consumers, many of us start to abandon a Web site when response times are more than 2 seconds. Some operational and most client-facing analytics have an informal 1 second SLA. Others talk about traditional BI, internal or operational analytics, and client-facing analytics, which is basically SaaS analytics for your customers. Some people talk about hot (1 second SLA), warm (10s of seconds), and cold (10s of seconds to minutes) analytics. I’ve had many conversations over the last 5 years because existing databases, in-memory technologies, or cloud data warehouses could not support what some call the 1 second SLA (Service Level Agreement). When to use Postgres, MySQL, in-memory databases, HTAP, or data warehouses ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |