Building software today requires a lot of dependency. Ten or 20 years ago, all of a company’s IT portfolio applications were at home in a data center; If you take a list of the company’s applications and services today, they are almost entirely in the cloud. In the past, if you wanted to make Enterprise Resource Planning (ERP) secure, you can simply go and check the log file to see who has access. But today’s software-as-a-service (SaaS) -developed world is very vague.
This is also true for small businesses – we employ about 200 people, but our teams use more than 100 SaaS products. As developers, integrating third party code into their workflows and adding variables quickly creates a software dependency fantasy.
Here we will look at how to register, prioritize and use the SaaS products used by your organization to help keep the entire digital supply chain secure.
First things first, self-assessment
As a Reporting SaaS trends From Blissfully, the average small business uses 102 different SaaS applications. Mid-market businesses averaged 137 applications, and enterprises averaged 288.
Taking inventory can be very difficult but it is important to keep running and producing enough manpower. The first step is to check with the accounting department to find out which SaaS subscriptions you pay each month. This is not a tag for any SaaS products you use for free, but it is a startup.
Once you know which SaaS products you are using, the next step is to make sure you have subscriptions. It often makes no sense to pay for two services that offer the same functionality – and it never makes sense to pay for something that has never been used, for example, a one-time service and one that has not been canceled.
Once your SaaS products are developed, you can prioritize the most important services in each category based on the importance or sensitivity of the relevant data assets – consider NetSuite or another ERP for Financial, Salesforce for customer details and so on.
Keep track of time information
A few mature SaaS services – I say approximately 10% – provide functionality to help secure your systems. But that doesn’t mean 90% of the time, organizations are on their own when it comes to improving security.
One of the best ways to monitor security is to model user behavior using time series data and look for abnormalities over time. Depending on an individual SaaS product or service, there may be five or more criteria for creating an accounting model that describes “normal” user behavior.
For example, for a developer forum, you can model commands such as “commit” or “clone” to understand the normal level of activity. Over time, they begin to see how often these orders are used on a daily, weekly, and monthly basis and come from geographical location. Suppose you have 80 engineers and almost all of them are based in the United States and Western Europe, but suddenly you see a connection that delivers orders from Ukraine. That could be something – and maybe – a clear red flag above.
Similarly, most organizations only perform a few clone operations daily or weekly; Hiring time series data to record activity over a period of a few months shows your typical use of your organization. Where you usually see three, you know you have a problem if your graph suddenly rises to 100 or more.
Remember that time series data modeling behavior does not prevent fraud, it helps teams respond quickly to unusual situations. Take it. Codekov violation From the beginning of this year – at the end of January – a malicious actor maligned in Codecov’s Bash Uploader script, but customers were not notified of the incident until April. If their teams were using time series data to model common behaviors, they would notice a fish simulation in a day or two, in contrast to the estimated two and a half. Months Codecov took action.
Finally, even though you know what to look for in each of the SaaS services you use, a common road block is getting the information you need to do this. That’s the key feature: I recommend that our teams find a SaaS solution – it allows you to programmatically expose logs through the API, use that data, and use machine learning to create your models. Avoid services that hide this basic ability beyond the basics.
Teams need to access log files for SaaS services that contain their most important data. In an industry where there is no violation, but time series data modeling can make the difference between responding quickly and allowing something to pass.