In the last year, we’ve seen the awareness of the need for data access control and security in cloud data warehouses pass an inflection point. Most companies we talk to now, especially in the FinServ and Pharma industries, know they must have it. We don’t have to convince them sensitive data needs to be protected in the cloud or show them stats about data breaches or regulatory fines. They get it. But how they decide to get to it is a different story. Some decide to go down the do-it-yourself or build-it-yourself route, but I’m here to explain why you shouldn’t.
Automation Greases the Wheels
Identity providers like Okta and Active Directory have done a great job of enabling companies to automatically generate as many users and roles in Snowflake as needed. Today admins can go from 0 users to 1000 in about an hour or two.
On the other side of the equation, ETL providers like Matillion, FiveTran and Talend have made it easy for companies to transport their data into Snowflake. In an hour or two, admins can move gigabytes or even terabytes of source data and have it ready and waiting for users to access.
These two forces come to a head at the intersection between them: connecting users with data and defining the relationships between them. How do you make sure only the right users have access to only the data they should have?
Enter BIY Data Access Controls
Many companies start with DIY or do-it-yourself: the trusty Snowflake admin or DBA decides to write a handful of SnowSQL Snowflake data access control policies, one at a time. This works when you have one or two new users a week requesting access. But chances are, if you’re using an identity provider to create your profiles, you’re already dealing in hundreds or even thousands of users. DIY just doesn’t cut it – doing that work can suck up hours or even days each week, bringing access for new users as well as any other data projects to a halt, not to mention the human errors that can be introduced. It simply won’t scale.
Okay, so then our ingenious database admin thinks, “I can BIY this” or build-it-yourself. “I have a tool that puts my users in automatically. And I have a tool that puts my data in automatically. I can fix this problem if I just spend the next week writing a tool that automatically connects these two domains together. Easy-peasy.”
But wait, let’s take a step back and think about this. Snowflake also gives admins a way to add users without an identity management tool and add their data without an ETL tool. So, what’s the advantage of using an Okta or Matillion? The answer is reliability, scale and automation – those software vendors have built solutions that save you time and just do it better.
Risk of Crossing the Streams – User x Data
It’s ironic that of the tools they could create on their own, some companies focus on connecting users with data. Obviously, they’re doing this because they haven’t yet found the Okta or Matillion to handle this. But the irony is that this is the most dangerous spot in the process – that intersection is actually where all the risks are.
You can add data to Snowflake, and it’s pretty safe when users can’t get to it. And you can add users to Snowflake, but they can’t do much without access to data. Very rarely do you get in trouble for adding a wrong user or the wrong data. If users aren’t connected to the data, the risk is near zero. It’s in the middle part where the streams cross that is fraught with risk. Connecting the wrong user with the wrong data can be very bad for a data engineer, data steward, or privacy owner.
You could BIY, but Are You Enterprise-Ready?
So, an admin can write a quick and dirty Snowflake masking policy, but can others read and work with it? Do you have a QA team to eliminate errors? Once you get a proof-of-concept to work on one or two databases, can you ensure it scales correctly and can run quickly across thousands? Do you have the time to integrate it with Okta or Matillion or Splunk? Do you have a roadmap that ensures it’s staying in sync with new private-preview Snowflake features, keeping up with your changing data and regulatory landscape, and addressing new user service needs? Can you ensure it actually works correctly – did you build in feedback and alerting on fails and errors?
In other words, do you want to hire 30 engineers and spend millions of dollars to build enterprise-ready software you can trust with the risky connection between users and data?
Automated Snowflake Data Access Control for the Win
Wouldn’t it just be easier to grab a third leg of your stool for data access controls to go with your user role and data transfer solutions? That’s where ALTR comes in. We’ve already invested the time and resources to build a world-class, reliable solution that automates and enforces the connection between users and data. It leverages all of Snowflake’s native data governance features while adding a no-code layer that makes it easy to apply and manage. It also shows you how users are accessing data to be confident that data is shared correctly. And because it’s SaaS, it’s fast to implement, starts at a low cost and can scale with your Snowflake usage – to hundreds of users and thousands of databases. (You could even think of it as Okta for Data.)