The Changing Nature of DDoS Attacks Demands a Rethink of Enterprise Protection

Most enterprises today have some form of DDoS protection. The changing pattern of attacks should prompt enterprises to re-examine how their DDoS protection is deployed, and whether that deployment is aligned with the way their networks actually operate

Keyur Jathal, Executive Director, Ishan Technologies

Consider a typical enterprise setup. One ISP provides internet connectivity along with its DDoS mitigation service. The organisation then adds a second ISP for redundancy, with a separate contract, a separate mitigation service and its own threshold for detecting attacks.

On paper, both links are protected. But there is an important gap: each provider is responsible for protecting the link it provides. Neither may have visibility into, or responsibility for, protecting the enterprise network as a whole.

This arrangement was largely designed around a different kind of DDoS attack. Traditionally, attacks involved large volumes of traffic overwhelming a link. The ISP would identify the attack, divert traffic to a scrubbing centre and initiate mitigation. In many cases, there was enough time for teams to coordinate and respond.

That assumption is becoming increasingly difficult to rely on.

The current attacks are short and relatively small in volume, lasting under 10 minutes and remaining below 1 Gbps, rose 208% year-on-year. Cloudflare’s Q3 2026 DDoS Threat Report recorded 18.2 million network-layer attacks during the quarter, averaging roughly 8,200 every hour. Of these, 96% ended in under 10 minutes and 98% stayed below 1 Gbps.

The implication is important. An attack can begin and end before an enterprise has even completed the process of escalating an incident across multiple service providers.

Smaller, Shorter and Harder to Distinguish

A DDoS attack works by directing traffic from multiple compromised devices towards a network, server or application, making it difficult for legitimate users to access the service.

The methods vary. Volumetric attacks consume available bandwidth. Protocol attacks exhaust network resources. Application-layer attacks are more difficult to distinguish because the requests can resemble normal customer activity.

Interestingly, the biggest attacks are not necessarily the most important indicator of where the threat is heading. The more significant shift may be towards attacks that are shorter, more distributed and harder to distinguish from legitimate traffic.

Cloudflare counted 251 attacks above 1 Tbps in Q3, down from 805 in Q2. However, NETSCOUT’s 1H 2026 report, covering more than 9.1 million attacks across 209 countries and territories, points to a different pattern. Attacks lasting a minute or less nearly doubled their share, from 1.6% to 3.1%. Carpet-bombing, where attack traffic is distributed across multiple IP addresses, also increased to 4.54% of attacks.

Another development worth watching is the use of residential proxy networks. These allow attackers to originate traffic from ordinary consumer IP addresses, making it harder to distinguish malicious traffic from legitimate users without inadvertently blocking genuine customers.

Taken together, these trends suggest that the challenge is no longer just about handling exceptionally large volumes of malicious traffic. It is increasingly about detecting attacks that are brief, distributed and designed to blend into normal network activity. This has important implications for how enterprises approach DDoS detection and mitigation.

The Gap Between Connectivity Redundancy and Network Protection

Having multiple internet connections is a sensible business decision. It reduces dependence on a single provider and improves connectivity resilience. But it also creates multiple entry points through which an attack can reach the enterprise network.

The challenge is that each ISP typically sees only the traffic passing through its own network. Its mitigation mechanisms are configured around the traffic patterns and thresholds of that particular link.

Consider an attack distributed across two internet connections. The traffic on each link may remain below the respective ISP’s detection threshold. However, when combined, the traffic could still overwhelm the server or application behind those connections.

Neither ISP is necessarily doing anything wrong. Each is monitoring and protecting what it has been contracted to protect. The problem is that neither has visibility of the complete picture.

This is not necessarily a technology failure. In many cases, it is a consequence of how enterprises have procured and structured their connectivity and security services over time.

Incident response can become equally complicated. Multiple providers mean different monitoring portals, escalation processes and definitions of what constitutes an attack. When an attack lasts only a few minutes, the time spent coordinating between providers can become a significant part of the problem.

It is worth clarifying that this concern is particularly relevant to organisations managing their own IP space and data centre infrastructure, as well as services such as payment APIs, VPNs and enterprise voice. Organisations whose websites sit behind a content delivery network may face a different set of challenges.

What to Check This Quarter

Enterprises should take a closer look at how their DDoS protection is structured. A useful starting point is to map every internet connection and understand what protection is associated with it. What are the detection thresholds? How quickly does mitigation begin? What does the service-level agreement actually commit the provider to?

More importantly, who has visibility of the combined traffic across all internet connections? If the answer is nobody, that is a gap worth addressing.

It is also important to examine whether mitigation can begin automatically or requires manual approval. Enterprises should understand who is accountable when an attack is distributed across multiple links. And rather than testing each connection independently, they should consider testing how the entire network responds to an attack across all links simultaneously.

Addressing this does not necessarily require a completely new security architecture. It calls for a more integrated approach to detection and mitigation.

This means having a common point of visibility across internet connections, automated mitigation mechanisms and traffic diversion arrangements established in advance across providers. These arrangements should be in place before an attack happens, rather than negotiated while an incident is underway.

The way enterprises think about connectivity resilience has evolved over the years. Having multiple links addressed an important question: what happens if one connection goes down?

Today, there is another question worth asking: what happens when an attack arrives across multiple connections simultaneously, potentially staying below the detection threshold of each individual link?

Having protection for every link is important. But ultimately, the objective should be to ensure that the enterprise network remains protected.

Authored by Keyur Jathal, Executive Director, Ishan Technologies