They’re no doubt betting people/companies that actually bother to ditch MS SQL are far fewer than customers they gain by offering Postgres and MySQL PaaS options. Obviously Microsoft didn’t bring MySQL and Postgres to Azure so people could ditch MS SQL, it was to bring people already using MySQL and Postgres to Azure. There is no reason the apps (.NET Core, one is on EF but we are ditching it) can’t run on Postgres, it’s just a matter of whether it’s worth the hassle of doing the migration. I will probably have to estimate the man-hour commitment for refactoring our existing apps and migrating tables and stored procedures over to Postgres, and it may be that it’s not worth doing based on my company’s financials and priorities, however my boss seems committed given the significant price difference. Microsoft will get their money one way or the other. Part of my job is and will be to be a leader in this effort.Īs far as we can tell, Azure SQL is so expensive because of the licensing- there is no free lunch and there is no getting away from the four or five-figure licensing costs companies incur for SQL Server, even in the cloud. We aren’t doing anything fancy or weird, it’s just tables with foreign keys and stored procedures. My company is looking at moving from SQL Server (Azure SQL really) to Postgres due to price differences. The 10k/month comparison isn't with a single DB instance, it's with several instances. If he is a DBA he shouldn't have to ask this question. If you're a DBA, moving to a different RDBMS is a big deal, but I doubt OP is at a company with DBAs otherwise they would be making this decision for him. Of course the tooling is different, but if you're coming from SSMS MySql Workbench or PgAdmin should be a godsend. You're describing the specifics with clustering but I doubt OP is doing that, I doubt it's a concern. "They are fundamentally different", okay how? What are the potential challenges? I mean I already mentioned SQL Server uses T-SQL and Postgres uses SQL so some refactoring has to be done there, but to me you're making out to be this significant thing when really for most apps I don't think it is. So far you have been making a lot of I guess blanket statements that don't seem to me to have any substance behind them. Btw, using your analogy of 80 hours to save $10k is great but the difference between Azure SQL and Azure Postgres is nothing like that unless you're comparing the top end of one and the absolute bottom end of the other. Look at the holistic need and identify your solutions based on that. This is why I said don't make the decision first on cost. Monitoring tools and the suite of tools you'd use to monitor, manage and secure the systems are different. Each of the major SQL solutions have a different performance impact when you have very large databases. MS-SQL clusters natively, meaning you can run multiple nodes active-active which Postgres (and MySQL) don't support without third party add-ons which each add complexity and potentially cost. INSERT statements are handled differently and those differences can have significant performance or correctness impacts at scale. They are both SQL servers but implement SQL support differently and in those differences are a lot of potential challenges. Your reply is fundamentally correct but clearly from a developers perspective. If you’re working on a GIS app, Postgres is the better choice, not that SQL Server cant store GeoJSON. Postgres has better support for GeoJSON and other spatial types. When I think of Postgres vs MSSQL, I of course look at what the development team is familiar with, because there’s ramp up time if they aren’t familiar with the new thing, but there’s hardly a learning curve going from SQL Server to Postgres. Moving to another RDBMS is not necessarily easy, but if it takes them 80 dev hours and saves the company $10,000/month, it’s probably worth doing and I can’t think of any drawbacks performance or otherwise. It’s all about the money, companies are all over decisions like OPs where one solution is significantly cheaper than the other. They both run on Windows, Mac and Linux and support C#, Python, Java, PHP, and anything else. They both use SQL (T-SQL is very close with only a few small nuances between how to handle say INSERTs with PK conflicts). They’re not fundamentally different- they’re literally both relational database management systems.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |