Explainer

DDoS protection for business websites explained

Learn how DDoS attacks hit business websites, how filtering and scrubbing work, and the practical steps to prepare without breaking normal traffic.

By the Cloudverg engineering team · · 6 min read

What it is

A DDoS attack, short for distributed denial of service, is an attempt to make your website or service unavailable by flooding it with more traffic than it can handle. The traffic comes from many machines at once, often a botnet of infected devices or a rented stress-test service. The goal is not usually to steal data. It is to make you unreachable, whether to hurt revenue, distract your team from another attack, or apply pressure during a dispute.

Attacks come in a few shapes. Volumetric attacks push raw bandwidth at you, like a huge wave against a small pipe. Protocol attacks exhaust connection tables on firewalls and load balancers using malformed packets or half-open connections. Application-layer attacks look like normal requests, just far too many of them, and they target expensive pages like search, login or checkout. A slow POST that trickles data for minutes can tie up a worker with almost no bandwidth.

You cannot stop a large attack at your own server. By the time packets reach your network link, the line is already full. That is why DDoS protection almost always sits in front of your infrastructure, at a provider with far more capacity than any single company needs. The job of that layer is to separate wanted traffic from junk before it ever touches your origin.

How it works

Most protection services work by changing where your domain points. You move DNS from your registrar or host to the protection provider, which gives you a new address that proxies traffic to your real server. Visitors resolve to the proxy, the proxy inspects each request, and clean traffic is forwarded. This is often called a reverse proxy or scrubbing setup. Your origin server should then only accept connections from the proxy, otherwise an attacker who learns its IP can bypass everything.

Filtering happens in layers. At the network edge, the provider drops traffic that matches known attack signatures and rate limits per source. For application traffic, it uses rules you define: block requests with odd headers, challenge suspicious clients with a JavaScript or cookie check, or show a CAPTCHA. Reputation feeds and threat intelligence flag addresses seen in previous attacks. Large providers also spread traffic across many data centers, so absorbing a flood becomes a routing problem rather than a capacity crisis.

You should tune rules to your own traffic. A rate limit that is fine for a brochure site will break an API that legitimately sends bursts. Health checks and failover matter too: if your primary data center is overwhelmed, DNS can point to a standby region. Logs from the proxy are the only place you will see the attack in detail, so make sure those logs reach your own monitoring. Many incidents are first noticed by a spike in 4xx or 5xx responses, not by an alarm from the provider.

Why it matters for a business

Downtime is not just a hosting bill. If your site takes orders, books appointments or runs a customer portal, every minute of an outage is lost revenue and support load. Staff may not know what is happening and will try to restart servers that are fine. Worse, an application-layer attack can exhaust database connections and leave data in an inconsistent state, so recovery takes longer than the attack itself.

The distraction is real. Small attacks are sometimes launched to pull attention while someone probes for weaker systems, like an exposed admin panel or an old VPN. If your team is firefighting a traffic flood, nobody is watching the login logs. Having a plan means the first hour goes to confirming the attack and following a checklist, not inventing one.

You also have a duty to your customers and partners. If your service is down, their operations stall, and they will ask what happened. A short, honest status page update during an incident does more for trust than a perfect post-mortem later. Knowing your baseline traffic and normal error rates is what lets you tell the difference between an attack and a bad deploy.

Common mistakes

The biggest mistake is leaving the origin IP exposed. Teams put a proxy in front of the site, then a subdomain like direct.example.com or an old DNS record still points straight at the server. An attacker finds it with a simple history lookup, and the protection is irrelevant. Audit every DNS record, certificate transparency log and outbound connection that might reveal the real address.

Another is testing only during a real attack. Run a controlled load test with your provider's approval, or use their test mode, and confirm that legitimate users can still log in and pay. Rules that block bots often block accessibility tools, older browsers and API clients. Check your mobile app and any partner integrations after every rule change.

People also forget the non-web services. Email, DNS and VPN endpoints get attacked too, and a DNS flood can take you offline even when your web tier is healthy. Keep your DNS hosted with a provider that has anycast and DDoS capacity, and do not put your only DNS server on the same network as your website. Finally, do not assume your cloud provider's default protection covers everything. Their network absorbs large floods, but application-layer attacks and cost spikes from autoscaling still land on your bill.

How to get started

Start by mapping what is public. List every domain, subdomain, IP and service that faces the internet, and mark which ones carry revenue or customer data. Find out what protection, if any, each one already has. Ask your host and cloud provider what they filter by default and where their responsibility ends. That gap is what you need to fill.

Then pick a layer. For most small and mid-sized businesses, a managed reverse proxy or CDN with DDoS protection covers web traffic well. Put it in front of the main site, lock the origin firewall to the proxy's address ranges, and set conservative rate limits. Add a status page and an incident runbook with names, phone numbers and the steps to enable stricter rules. Decide in advance who can authorize blocking a country or turning on a challenge for all visitors.

Practice before you need it. Run the runbook once with a fake alert, time how long each step takes, and fix the parts that rely on one person. Review the rules quarterly, since your traffic changes. If you want an outside view of your cloud setup first, a structured review can help: Cloudverg offers a Cloud security review for $6,500/month, a fixed-scope read-only check of up to five cloud accounts that ranks exposure and logging gaps with the exact fix for each. It is one option among others, and it pairs well with a DDoS plan rather than replacing one.

Frequently asked questions

Can I stop a DDoS attack with a firewall?

A firewall helps with small attacks and bad packets, but it cannot absorb a flood that fills your internet link. Large volumetric attacks need filtering upstream at a provider with more capacity. Use both: upstream protection for volume, local rules for application behavior.

Will DDoS protection slow down my website?

A well-configured proxy adds little latency, often a few milliseconds, and can speed things up through caching. The risk is over-aggressive challenges that make real users solve puzzles. Test logins, checkout and API clients after each rule change.

How do I know if I am under attack?

Watch for sudden drops in availability, spikes in requests from odd locations, and rising 4xx or 5xx errors. Slow response times on dynamic pages with normal bandwidth often point to an application-layer attack. Compare against your normal baseline before assuming the worst.

Monthly plan

Want help with this? Cloud security review

$6,500/month

2 to 3 weeks · Starts within 10 business days

A read-only review of identity, network exposure, encryption and logging across your cloud accounts, with a prioritized fix list.

Renews monthly. Cancel any time.
What is included

Related guides

More cloud security guides   Security and compliance

Smallest way to start

Get a senior architect on your problem this week

Ninety minutes a month on the cloud, AI or security question that is blocking you, a pre-read of your material and a written summary within two business days.

$450/month

Renews monthly. Cancel any time.