Securing the Data Pipeline

Businesses are using a vast network of digital pipelines for their data, but many are doing so with little thought about the security risk. Find out how can they can be protected against an increasingly uncertain and volatile threat landscape.

By Chris Cicotte, Security Messaging Lead, Dell Technologies

While experts often contend that data is the new oil, they usually ignore the corollary: oil needs a pipeline. In 2020, we’re already using a vast network of digital pipelines for our data, but many companies are doing so with little thought about the security risk. What happens if we mismanage our data pipeline, and how can we be sure we don’t?

COVID-19 taught us that we’re susceptible to fast-spreading physical viruses, but the World Economic Forum has warned that we’re at just as much risk from digital ones. It predicts a global cyber pandemic, in which fast-spreading software pathogens would incapacitate millions of devices, resulting in a possible cyber-lockdown as we try to recover.

In times of social crisis, our focus has traditionally been on keeping crucial pipelines running. Energy pipelines keep our homes warm, our businesses running, and our cities illuminated. Food pipelines keep us fed. In 2020, we must also consider the data pipeline.

Each day that the world was offline would cost over $51 billion—but it isn’t just the cost that would cripple us. The world relies on data connectivity for everything from finance to public health. If the data pipeline failed, critical social functions could fail with it. It’s therefore critical that we keep that data pipeline both open and secure to enable organizations to not only survive but also to be bold, entrepreneurial, and unencumbered by the growing wave of digital threats.

The data pipeline is more complex and more fluid than ever before, thanks to the rise of cloud-based data integration. The data traveling along this pipeline can easily become toxic. It often represents highly sensitive personal information which is subject to strict regulations such as Europe’s General Data Protection Regulation and the California Consumer Privacy Act. Mishandling it can cost organizations dearly. As we struggle to control our data, its volume is growing, which means that there’s more of it to manage.

A New World of Zero-Trust Architecture

This vast data pipeline has rendered yesterday’s cybersecurity protections untenable. Years ago, we relied on a hard perimeter of firewalls that protected data inside our corporate networks. Today, data travels worldwide between our on-premises systems, the cloud, business partners, and remote workers. Increasingly, the applications that access this data are ephemeral, spinning up in containers that are destroyed minutes later. The traditional perimeter was never designed for this. It has disappeared to be replaced by a new security model: zero trust.

A zero-trust architecture focuses security on what matters most: the data traveling along this intricate pipeline. Instead of trusting specific devices or IP addresses, it forces all users and devices to authenticate themselves when accessing sensitive data and applications, wherever they are and whichever device they’re using. A zero-trust network also focuses heavily on the concept of least privilege. It always grants only the access privileges to data and applications that a particular session allows and no more.

Zero-trust technology doesn’t just operate at the network layer: It has expanded to the application layer, taking parameters from both elements of the technology stack into account. Along with the user’s identification, it examines factors including the user’s role and responsibility, the network that they’re accessing from, the resource that they’re requesting, and what they want to do with it. These parameters change frequently, so it requires repeated authentication. In some cases, the user might not be human at all; they could be among millions of IoT devices in far-flung physical locations vulnerable to physical tampering.

By making identity a key part of the authentication process, zero-trust architecture recasts identity as the new perimeter. It foregrounds identity and access management as a key tool in authentication and authorization. That also sharpens the need to protect that identity. Multi-factor authentication is a key part of the verification process in a zero-trust world because it helps avoid account hijacking.

Building a Zero-Trust Architecture

Building a zero-trust architecture is a wide-ranging, multi-disciplinary initiative that reframes the core cybersecurity principles upon which companies have relied for years. It also touches all parts of an organization’s technology stack. It will be difficult to retrofit existing technology architectures with this overnight.

Companies should think about zero-trust initiatives in the long term. The transition will be gradual as teams build out solutions incrementally. IT teams should make new product or service acquisitions with zero-trust architectures in mind.

This new approach to data requires more than just technical input. We must also rethink our data governance principles. We must focus more closely on the provenance of the data flowing along our pipelines, along with where and how it is stored when it reaches its destination, and for how long.

Who owns this data? Who has access to it during its journey? What did they do to it, and why? Answering these questions should involve a broad array of stakeholders, from legal and compliance to technical experts.

The security team will play a key role in these conversations, and this requires a fundamental rethink of the security team’s role in the system design process. Traditionally, security was the last group to be consulted about technical decisions. It was seen as the “office of no,” quashing attempts at innovation with its strict resistance to change.

Today, the data we’re trying to protect is so bound together with privacy and governance concerns that we cannot talk about it without acknowledging and pursuing security. As the gap between security and system architecture becomes increasingly airtight, our perceptions of security must change. The security team must have a seat at the table.

Reframing our security model and rebuilding these internal relationships will help protect our data pipelines against an increasingly uncertain and volatile threat landscape. We owe it to society to make this a priority.