Even before lockdowns, there was a steady migration toward more flexible workforce arrangements. Given the new normal of so many more people working from home—on top of a pile of evidence showing that productivity and quality of life typically go up with remote work—it is inevitable that many more companies will continue to offer those arrangements even as stay-at-home orders are lifted.
Unfortunately, a boom in remote access goes hand-in-hand with an increased risk to sensitive information. Verizon reports that 30 percent of recent data breaches were a direct result of the move to web applications and services.
Data is much harder to track, govern, and protect when it lives inside a cloud. In large part, these threats are associated with internet-exposed storage.
Emerging threat matrix
Traditionally, system administrators rely on perimeter security to stop outside intruders, yet even the most conscientious are exposed after a single missed or delayed update. Beyond that, insiders are widely considered the biggest threat to data security.
Misconfiguration accounts for the vast majority of insider errors. It is usually the result of failure to properly secure cloud storage or firewall settings, and largely relates to unsecured databases or file storage that are directly exposed on a cloud service.
In many cases, employees mislabel private documents by setting storage privileges to public. According to the Verizon report, among financial services and insurance firms, this is now the second most common type of misconfiguration error.
Addressing this usually means getting open sharing under control, figuring out where sensitive data resides and who owns it, and running a certificate program to align data access with organizational needs.
Optimistically, companies hope that a combination of technological safeguards and diligence on the part of users—whether employees, partners, or customers—will eliminate, or at least minimize, costly mistakes.
Other internal threats come as a part of a cloud migration or backup process, where a system admin or DBA will often stand up an instance of data on a cloud platform but fail to put inconvenient but necessary access controls in place.
Consider the example of cloud data warehouses. Providers such as Amazon, Google, and Snowflake now make it simple to store vast quantities of data cheaply, to migrate data easily, and to scale up or down at will. Little wonder that these services are growing so quickly.
Yet even the best services need some help when it comes to tracking data access. Some tools makes it easy to authenticate remote users before letting them inside the gate of the cloud data warehouse. After that, though, things often get murky. Who is accessing which data, how much of it, when, and from where?
These are issues that every company must confront. That data is ripe for exploitation by dishonest insiders, or by careless employees, with serious consequences. In more fortunate circumstances, it is discovered by security teams, or by management who make an irate call to the CISO.
Born in the cloud
More approaches to data security that are born in the cloud are now appearing, and the new normal means the enterprise is motivated to adapt. As most organizations turn to the cloud for what used to be on-premises IT deployments, the responsibility and techniques to secure the infrastructure and applications that hold data are also being moved to the cloud.
For instance, infrastructure-as-a-service (IaaS) provides virtualized computing resources like virtual firewalls and network security hardware, and virtual intrusion detection and prevention, but these are an intermediate step at best.
The idea is that IaaS can offer a set of defenses at scale for all of a cloud provider’s customers, built into the platform itself, which will relieve an individual cloud customer from having to do many of the things that used to be on-premises data-protection requirements.
But what has really changed? A top certification may be enough to be called “above average” data security, but in reality that security still remains totally contingent on perimeter defenses, hardware appliances, and proper configurations by system administrators and DBMs. And it’s still only as good as the data hygiene of end users. There are a lot of “ifs” and “buts,” which is nothing new.
Data Security-as-a-Service (DSaaS) complements IaaS as it integrates data protection at the application layer. This places data access services in the path between users who want data and the data itself. It is also portable because it goes where the application goes.
Developers can embed data access governance and protection into applications through a thin layer of technology wrapped around database drivers or APIs, which all applications use to connect to their databases. An obvious advantage is that this is more easily maintained over time.
Data security is a shared responsibility among security pros, end users, and cloud providers. As the new normal becomes reality, shared responsibility means that a cloud provider handles the underlying network security such that the cloud infrastructure ensures basic, customer-level network isolation and secure physical routers and switches.
From here, under the DSaaS model the cloud service provider offers DSaaS—or else the customer provisions it through a third party—as a set of automated data security components that complete a secure cloud environment.
This makes it possible to govern each user at a granular level so that they access only the types of data they should, and perform only those actions with the data for which they are authorized. CISOs can implement and adapt rulesets to govern the flow of data by type and role. In terms of data protection, application-layer data security makes it possible to isolate and block bad traffic, including excessive data volumes, down to an individual user.
From this perspective, DSaaS can act as both an intrusion detection system (IDS) and intrusion prevention system (IPS). It can inspect data access and analyze it for intrusion attempts or vulnerabilities in workload components that could potentially exploit a cloud environment, and then automatically stop data access in progress until system admins can look into the situation.
At this level it is also feasible to log data activity such as what each user does with the data they access, satisfying both security and compliance—a notable accomplishment, considering that the two functions are often at odds with one another.
Incorporating security at the application layer also offers data protection capabilities that are similar to network intrusion appliances, or security agents that reside at the OS level on a virtual machine or at the hypervisor level.
Moreover, DSaaS governance and protection is so fine-grained that it does not inhibit traffic flow, data availability, and uptime even in the face of multiple sustained attacks.
Everyone is talking about how the “new normal” is impacting data security, but the enterprise was well on this path before the pandemic. It is tempting for vigilance to give rise to pessimism since data security has too often been a laggard, and an inventory of the cloud data-security bona fides of most companies is not encouraging.
However, data protection and governance can be assured should we adopt shared models for responsibility and finely tuned, application-level controls. It’s a new world and we can be ready for it.