![]() ![]() I do want to caveat this, however, that with like any cloud tech out there, I'm pretty certain you could set yourself up in a way that inadvertently optimizes for cost (in a bad way), and destroy your budget accordingly. I haven't encountered a single use case so far where the cost delta of using Snowflake vs a native solution came anywhere near what you'd pay for even a single engineer's salary. So, to answer your specific question (is Snowflake so expensive that its cost dwarfs the cost of an engineering team maintaining basic services), I'll answer with an enthusiastic no. Cost is one of the questions that come up the most often. Would love to answer questions or debate the above. I'm also the co-founder of one of the leading Snowflake consulting firms in North America. I wrote a fairly extensive article on that very topic a few years ago and I'm quite passionate about "the right way" to do multi-cloud strategies overall. Take also a look at Wasabi, they're building an S3 competitor (not AWS as a whole, only S3) that's more cost competitive than S3 and can be used through AWS' own command line tool as Wasabi's API is bit-compatible with it (meaning that migrating to their service is as easy as changing a single configuration line in the command line tool).Īnyway, I digress. Competitors will instead focus on more specific verticals/use cases and go much deeper to try and differentiate themselves to a point where they yield the type of ROI warranting their choice over a native service. The big three (AWS, Azure and GCP) have cornered the general compute market and significantly elevated the barrier of entry to the cloud market. That's a pattern you'll see more and more of, I believe. Also, Snowflake runs on AWS under the hood and is natively integrated to S3 for storage (and you can even provision it through the AWS marketplace), so the line between native AWS service and external cloud service provider is much thinner than for other tools. They came out with serverless capabilities for Redshift (in big part to try and make it more competitive with Snowflake, which has separated compute and storage layers) but it's still fairly limited and expensive. Still using Redshift as a case study, I'd say that it's fairly high-maintenance for a managed service (you still have to manage your clustering, scale your hosts out, etc), where Snowflake is much more hands-free. ![]() Reality is, sometimes, AWS services (Redshift is a good example of that) simply aren't as good as non-AWS tools, to a point where using a non-AWS option has a high enough ROI to justify the (sometimes relatively minor, sometimes significant) overhead of using a non-natively integrated solution. ![]() I understand where you're coming from, I used to be an AWS architect and thought similarly to you for a long while, until I encountered a few situations that changed my perspective. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |