Press "Enter" to skip to content

Category: Security

Protecting ML Models and IP

Published 2022-05-18 by Kevin Feasel

Pete Warden has some advice:

Over the last decade I’ve helped hundreds of product teams ship ML-based products, inside and outside of Google, and one of the most frequent questions I got was “How do I protect my models?”. This usually came from executives, and digging deeper it became clear they were most worried about competitors gaining an advantage from what we released. This worry is completely understandable, because modern machine learning has become essential for many applications so quickly that best practices haven’t had time to settle and spread. The answers are complex and depend to some extent on your exact threat models, but if you want a summary of the advice I usually give it boils down to:

– Treat your training data like you do your traditional source code.

-Treat your model files like compiled executables.

Read on to see why Pete came to this as the appropriate answer, as well as what I have to consider a sly mention of duck boat tours.

Comments closed

Thoughts on Code Obfuscation

Published 2022-05-11 by Kevin Feasel

Joy George Kunjikkur reminds us of worse times:

Long long ago I was given the special task of hiding code. Hiding code..what? Yes, we have to deliver code in such a way nobody should be able to reverse engineer.

I’ve run into problems around this in modern code as well. For example, using client-side React means that you aren’t going to be able to hide secrets like credentials, connection strings, etc. in a way that users absolutely won’t be able to see them. In the SQL Server world, some companies use encrypted stored procedures, which is a joke considering that you also need to ship the keys to decrypt those procedures, meaning that an enterprising user can get around your obfuscation attempt in moments.

Comments closed

Azure SQL DB ARM Template Conflicts with Azure AD Administration

Published 2022-05-06 by Kevin Feasel

Joao Antunes points out a potential timing issue around combining Azure Active Directory administration with Azure SQL Database ARM templates:

ARM templates are widely used when we need to repeatedly deploy solutions/infrastructures in the cloud. Leveraging the concept of infrastructure as code ARM templates are a powerful resource to ease our daily job, however we might face some challenges when using them.

When we are creating several resources within the same template – using Json or Bicep – it’s crucial to make sure that all resources are created in the right order, ensuring that all depending on resources are fully provisioned before you move to the next operation.

Error (internal server errors) and conflicts  can occur during our ARM template deployment and it could be difficult to troubleshoot or understand the root cause of them.

Read on for one annoying error and its fix.

Comments closed

Transparent Data Encryption in RDS

Published 2022-05-04 by Kevin Feasel

Tom Collins enables Transparent Data Encryption in AWS’s RDS:

Does AWS RDS SQL Server support built-in SQL Server Transparent Data Encryption (TDE)?  Yes , is the short answer , but there are some limitations  . 

Before I dive into the AWS SQL Server RDS TDW limitations , in this post  we’re discussing SQL Server TDE as opposed to RDS Encryption covering Aurora, MySQL, MariaDB, PostgreSQL, Oracle, SQL Server

Read on for the limitations Tom mentions.

Comments closed

Securing Azure Storage

Published 2022-05-03 by Kevin Feasel

Craig Porteous continues a series on Azure Data Platform security:

This is the third in a series where I look at all of the resources common to a Data Lakehouse platform architecture and what you need to think about to get it past your security team.

Building upon Azure Databricks, I’ll move from the compute engine to our blob and data lake storage. Things are a little simpler to secure but the plethora of options available can have significant impacts on usability and cost so it’s important to understand the impact before baking them into your design.

Read on for some good advice around securing Azure storage accounts.

Comments closed

S3 and Redshift Data Movement with Role Chaining

Published 2022-04-29 by Kevin Feasel

Sudipta Mitra, et al, talk AWS security:

This post presents an approach that you can apply at scale to achieve fine-grained access controls to resources in S3 buckets and Amazon Redshift schemas for tenants, including groups of users belonging to the same business unit down to the individual user level. This solution provides tenant isolation and data security. In this approach, we use the bridge model to store data and control access for each tenant at the individual schema level in the same Amazon Redshift database. We utilize ASSUMEROLE and role chaining to provide fine-grained access control when data is being copied and unloaded between Amazon Redshift and Amazon S3, so the data flows within each tenant’s namespace. Role chaining also streamlines the new tenant onboarding process.

Read on for an overview and tutorial.

Comments closed

A Power BI Report for Power BI Report Access

Published 2022-04-28 by Kevin Feasel

Gilbert Quevauvilliers sets up an infinite loop:

In this blog post I show the final part which is how I created the Power BI report which takes the previous 3 steps and then creates the Power BI Report.

I am going to show you how I got the data in using Power Query and then created the Power BI report.

Read on for the process. But now I want a report to see who has access to the report for who has access to reports. And I think I need a report for that layer. And that layer. And…

(Shh, yes, I know you can get that all from the same report but it’s so rare I get to make a “Turtles the whole way down” reference).

Comments closed

Azure Databricks Security Considerations

Published 2022-04-26 by Kevin Feasel

Craig Porteous provides some advice on configuring Azure Databricks:

Azure Databricks is an analytics platform and often serves as the central compute component of a data platform, to process ETL/ELT data pipelines and data science workloads. As Databricks is a third-party platform-as-a-service offering securing it works differently to most other first-party services in Azure; for example, we can’t use private endpoints. (More on these in the Azure Storage post)

The two main approaches to working with Databricks in our secure platform are VNet Peering or VNet Injection

Click through to learn the difference between these two, as well as a few other factors to keep in mind as you’re deploying Databricks.

Comments closed

Keeping Secrets in Azure DevOps

Published 2022-04-25 by Kevin Feasel

Kevin Chant has a secret:

In this post I want to cover how you can keep your Azure Synapse secrets secret in Azure DevOps. Because you need to do this if you are working with production deployments.

With this in mind, I want to raise more awareness about it and make sure others avoid putting secrets directly in their pipelines like in the below example.

Read on to understand what options are available to you. My preference involves Key Vault references but there are alternatives available.

Comments closed