What’s so great about Snowflake? (non-technical)

Non-Technical Demystification of The Largest Software IPO of all Time

The hype train is certainly pulling into the station. You may have heard about Snowflake, the software company that Warren Buffett invested in and is now having its Initial Public Offering (IPO) So really, what exactly is so great about Snowflake (in non-technical terms)?

The meter turns off

Sitting in standstill traffic with the taxi meter running is a very special feeling. You are going nowhere and getting poorer all the while. Well Snowflake turns the meter off when data is not moving. Petabytes of data are out there, pretty much just sitting there unused. Sometimes for years. With older data stores, the meter is always running for capacity to call and crunch the data. Not with Snowflake. If the data is just sitting there, the processing capacity goes to sleep and the meter turns off. For big companies with lots of data this is huge. The Snowflake for “data just sitting there” is a very respectable $23 per terabyte per month. (That’s about enough to store over 600 full length movies).

If we compare that to something else, like Amazon AWSRedshift, we find that the meter does not shut off. So the cost per terabyte is about $306 per month. It’s not an apples-to-apples comparison but for data that “mostly just sits there,” $23 v. $306 is an eyeopener. And, frankly, even AWS Redshift is eye-openingly cheap compared to corporate “on prem” storage costs, which can get into the thousands of dollars per month.

So, if the meter is off a lot of the time, what does this mean to performance when I need it? Actually, the way Snowflake is set up, performance is cleverly optimized. Even as the data size increases Snowflake distributes performance automatically behind the scenes. This relieves a bunch of setup and engineering headaches.

Digital sweatpants to formal dress code

What are digital sweatpants? Come on over, exactly as you are. Snowflake grabbed an idea from Google Big Query that made it really popular. Just jam data into Snowflake as it is. Don’t worry about the structure. Old databases require a strict dress code to be saved in the database. It takes some work.  Snowflake? The data can be saved exactly as it is. (There are substantial challenges with this) but if you are in a pinch and just want the data saved without a lot of headache…Snowflake is your most comfy pair of sweatpants.

But there is more. Want to dress up like an “old fashioned” (relational) database. Yep. Snowflake can do that too. So it tackles structured (strict dress code) and unstructured (no dress code). (We will have a follow up post about structured and unstructured ETL for Snowflake using K3.)

That friend that is friends with so many friends

Snowflake runs on all the big cloud providers. This is kind of a big deal. Some people are really loyal to AWS and others to Azure. Google…sure. Oracle Cloud? Ummmm….Snowflake is agnostic to all this so it can be used by everyone wherever they like. This is really important because as companies make the leap to cloud they tend to focus on just one over time.

And if SQL does not mean anything to you…don’t worry. SQL (pronounced seequil) is the historical baseline database language. There has been a wave of new databases that do not use SQL. It ruffled some feathers because people always struggle with new languages. Snowflake? Straight SQL and voila everyone is happy.

What’s it all mean?

Companies still need to keep plenty of important historical data. In Snowflake, it’s all on tap and ready to go, but without the meter running. Customers storing data are getting a really nice discount by using Snowflake.

But, think about this. Let’s say your business has a factory manufacturing cars. That factory churns out lots and lots of data. Right now it’s too much, too hard to store. Even if we did save it, we would not know what to do with it. Say, here’s an idea. Put it into Snowflake. You are giving your company’s future self a high five. We don’t know what we want the data for as of now, but in the future? Who knows. At $23/TB/Mo that’s a pretty cheap option. Maybe you don’t see it now, but in five years it could be your promotion.

What now?

Ready for your off-the-shelf adaptor to Snowflake? K3 is used by some of the world’s largest companies to pick up data from anywhere in your organization, intuitively transform it and get it to where you need it, including Snowflake! Everything in K3 is streaming and real-time. We’d love to show you how you can get the most out of Snowflake.

Always happy to show our stuff.