Radware Software and Solutions

Radware offers a cloud, software, and hardware-based solution. It deals with all three platforms. 

1. They have a hardware device on which their software can be installed. We can manage all the load balancing with it. 

2. Similarly, for the Radware software, we can install the OVA file on our server and configure all the admin backend servers on it to perform services. 

3. In the cloud, we can use their API service to create a virtual platform for clients on which they can deploy and run their applications.

Cloud WAF blocks unknown threats and attacks. We have a monitoring tool, and security patches are released monthly. We can deploy these signatures on the WAF, which identifies threats based on IPs. There are multiple signatures for various attacks, like bot attacks, that we can monitor.

There is a forensic dashboard where we can identify real-time events, hits, and blocks. If there are genuine requests being blocked, we can deploy a custom page with a case number for users to resolve issues. For example, if a user triggers the Web Application Firewall (WAF) due to a misinterpreted service, they will see a blocking page with a case number. There's also an option to refine the WAF settings if it blocks a genuine request.

I also work with the API discovery feature in the Cloud solution. Once the API is enabled and the application vendor provides the API key, we can deploy our application. If the API is correct, it functions properly; otherwise, issues are highlighted on the dashboard. For example, cross-site scripting is blocked at the label level.

API discovery is straightforward to use. There is an option to add the API stream. If the API is correct, it will be processed; otherwise, the API service is blocked.

The dashboard provides multiple features and analytics tools to identify API issues. If there is a cost issue with an API, it can be identified, and we can report it.

It's not difficult to work with the API discovery feature because everything is reflected on the forensic dashboard. There's an option within the dashboard, under the security section, where you define the correct API. You can also identify and exclude specific APIs if needed. There's only one option to add to the API stream. If the API is correct, it will be processed; otherwise, it's blocked.

It's not difficult to identify API issues because when we define the API call, and it is incorrect or not valid, it won't sync with the vendor's application. They identify this and generate a blocking request, which helps us easily identify the issue.

It's mostly for the Alteon service. The Alteon load balancing part, particularly the SSL offloading and WAF offloading, is crucial. Offloading allows us to monitor and identify issues easily. I believe the SSL offloading is the most valuable feature.

It's easy to use, and the configurations are similar across different vendors. Compared to F5 and Citrix, Radware is easier to communicate with and use. The configuration process is simple, involving the creation of groups and pools, much like in F5. The SSL offloading is also very easy. Overall, I think it's a good solution.

The service we use through the cloud is very easy. We have one dashboard to manage everything, which is convenient.

The analytic dashboard could be integrated with other platforms like Splunk. In Splunk, the dashboard shows multiple things, and I think Radware could improve its dashboard in that regard.

In the WAF part, there are multiple things that are initiated, such as updates and patches. There's a global issue right now that we need to monitor on our side. I think the ability to monitor server-level updates and patches should be integrated into the WAF.

I have been using it for three to four years. 

If we raise an issue, they usually identify and resolve it by the end of the day or the next day. There haven't been any escalated cases on the cloud. However, we did encounter one issue regarding the filter and signature. We created a policy to block access from Pakistan, Bangladesh, and other specific locations.

Although the policy was in place and checked, users from those locations were still able to access the application. This was a bug that we reported to the technical team. They identified an issue with their software version and provided us with a new version to update. After the update, the blocking feature worked correctly.

It's scalable. We can customize it as per our requirements. We can customize it in most cases.

In the State Bank of India project, we deployed it, and I believe two or three other banks are using Radware's WAF. Some applications are deployed globally, meaning they're used in Australia, America, and multiple countries. We have multiple deployment options for that. For example, the YONO application is deployed globally and used by many users in different countries. We can easily identify and track that traffic on the dashboard.

In addition, they have also deployed the DDoS service in WAF. So, in case of a DDoS attack or something similar, they can easily identify and monitor it.

It's software-only, so most of the time it works as intended. However, I did raise one request about a filter option in the dashboard not working perfectly. We identified that there was a version issue, and they fixed it in a new patch. They were able to easily identify and resolve the version issue.

They are not globally available but can manage and support us within a range of five to seven. They can usually provide support easily.

I have worked on Citrix as well. Both Citrix and Radware are similar, but in Citrix, some things are more lengthy. Radware is better integrated and easier to understand, so anyone can use it.

I haven't directly compared them, but F5 is very popular globally. Both are similar, but Radware lags behind F5 in a few features.

In terms of user experience and management, Radware is easier. However, F5 has better performance. 

Both are cost-effective, but Radware is less expensive because F5 licenses are costly. Technically, Radware is easier to understand.  

We currently use the integrated WAF option on the same device in our application (SBI). There is no dedicated WAF solution. There are two options: license-based and integrated. Using the integrated part helps to identify blocks and other issues effectively.

Integrating with other systems and applications in the environment:

Integration is not difficult. In the dashboard, under the policy section, we can find virtual services and easily enable the API service. Once enabled, WAF monitoring should also be enabled. We can then identify the application's requirements, like JSON ID, cookies, headers, what should be whitelisted, body size, etc. 

We can gather this information from the application owner during deployment to determine what needs to be whitelisted, such as extensions, zip files, XML files, and cookies.

Once we deploy an application, it doesn't take too much time because the application is already deployed. We also use the load balancing feature, so we just need to enable the security web application service. There is an option for this under virtual services where we can also enable it for bot protection. I think anyone can easily manage it if they know about these things.

Radware is signature-based. The patches and signatures are important because we cannot easily monitor them ourselves. They are regularly updated, I think, weekly, so that's helpful. I think this regular update makes it easier for us.

We are a managed service provider (MSP) for Radware. The technical support is handled directly by Radware, but we manage the technical aspects.

We use integrated and cloud solutions because we manage multiple applications for multiple vendors. Some vendors are using the integrated WAF, which is good. The cloud part is also managed by us, not the customer. We deploy everything, including signatures and patches, if needed.

We can deploy it within a month. It's very easy to deploy and work with. If you create load balancing and WAF configurations, both are very simple. The deployment process is easy if you know how to configure it. Anyone who knows the basics of networking and security can easily deploy it. The dashboard and management are also simple. There is no confusion.

If you're creating a virtual service, you can easily create the virtual service port and configure the backend server. It's very simple.

In F5, when creating a group, you need to take one pool service. But in Radware, you can create one group and easily select it. The dashboard and configuration in Radware are very simple.

We mostly deploy in one-arm mode, but there's also a two-arm mode. In one-arm mode, all applications and servers are on the same subnet. We take a single IP from the subnet (e.g., 10.86.11.x). We need three IPs: one for management and two for deployment and virtual services. 

When deploying an application, we can use the same IP range. We deploy all backend servers on the virtual service. We select the backend servers and multiple ports based on the requirements. We then configure the services on the virtual service and review everything. For networking, we need to do NATing if the application is globally accessible, which is also very simple.

The dashboard and conciliation aspects are straightforward in Radware.

Just as in the same domain, we can deploy mainly in one-arm mode, or two-arm mode. There are two different modes. Okay? But typically, we deploy in one-arm mode. In this mode, all applications and services should be on the same subnet. We can take a single IP from the same subnet, for example, if you have a subnet of twenty-three, like 10.86.11.something. We require three IPs in total. One IP for management, and two others for deployment and the virtual service. If we deploy an application, we can use that same IP range. We take it. And on this virtual service, we can deploy to all back-end servers. We can select the back-end server and multiple ports based on the requirements. We can select the ports we need and configure the services on them. On the virtual service, we can configure all the services and review everything. For networking, we need a NATing part if the application is globally accessible so we can NAT through their public IP. It's a very simple deployment process.

For a new project, it might take longer than a month due to approvals and networking configurations. These processes, especially to get approvals for NATing and network paths, can be take time. That's why it takes almost two months. However, if everything is ready, deployment and testing can be completed within five to ten days.

Two resources are enough for the deployment. From a maintenance perspective, not much is needed.

It does bring ROI.

Radware is less expensive because F5 licenses are costly. F5 charges for each and every license. For every virtual service you create, you need to pay additional license fees. The licenses are more COSTLY compared to Radware. 

Radware also has lower annual maintenance costs (AMC) compared to F5. F5 is more expensive than Radware, but it's the leading product globally.

It's not very costly because everything is license-based, all things depend on the license and annual maintenance contract (AMC). If you have an AMC, the cost will be higher. Without an AMC, the cost is less because the product itself is less expensive. But if you have the AMC, the cost will be higher.

If companies provide the signatures and patches perfectly because we can't easily identify new viruses or threats, we rely on the solution company to regularly update their software and devices. Radware is one such company that updates its patches and signatures monthly. They allow us to review all the CVEs and update their patches accordingly. So, I think it's a good option.

Overall, I would rate it a seven out of ten because there are some issues in the cloud part, where it lags. 

Radware AI Security Suite protects our AI-powered application in Microsoft Copilot's deployments from AI-specific threats, ensuring our governance, visibility, and real-time security monitoring.

Radware AI Security Suite helps with governance, visibility, and real-time security monitoring in our Copilot deployment by providing us a centralized view of AI applications and tools, along with user interaction across our environment, helping us understand which AI services are being used and how they are connected.

Our main use case is to continuously monitor AI activity and identify risky behavior or misconfiguration, and Radware AI Security Suite is helping us with that.

Several internal and external factors prompted our organization to explore solutions like Radware's Agentic AI Protection, including increased adoption of AI technology, such as Microsoft Copilot and other AI-powered applications, creating a need for stronger governance, visibility, and security controls, as well as the awareness of AI-specific threats such as prompt injection, jailbreaking, and unauthorized tool uses highlighting gaps that necessitated a protective security approach.

The main risk area my organization prioritized when considering Radware's Agentic AI Protection is data leakage, as protecting sensitive data has become a key requirement due to increased use of AI tools including Copilot and other AI agents' applications.

The most critical LLM-related risks my organization prioritized addressing with Radware's LLM Firewall include prompt injection attacks and sensitive data leakage, as AI agents have become a crucial part of our organization, necessitating a data control tool to prevent exposing confidential business data.

From the security perspective, Radware AI Security Suite positively impacts our organization by providing real-time monitoring and behavioral analytics capabilities that help us identify suspicious activities, prompt injection attacks, unauthorized access patterns, and abnormal AI agents, improving our ability to detect and respond to potential threats before they can affect our business.

One of the most significant benefits since implementing Radware AI Security Suite has been faster detection and response to AI-related security events, as the behavioral analytics and real-time monitoring capabilities help our security team identify anomalous activity, such as suspicious prompts and unusual access patterns, facilitating earlier investigations.

Upon activating Radware's Agentic AI Protection, the number of agents identified in our environment compared to prior estimates showed that while we had visibility into officially approved AI tools, Radware's solution helped identify additional agent integrations and AI-driven workflows being used across different teams.

Radware's Agentic AI Protection revealed unexpected insights about our internal network topology, including previously undocumented AI-related connections and workflows in our environment.

The best feature Radware AI Security Suite offers is AI agent discovery and visibility, along with behavioral analytics, LLM firewall capabilities, and real-time runtime protection, providing a centralized view of AI agent tools and interactions, making it easier to govern and secure AI deployments.

The behavioral analytics capability is one of the most valuable aspects in this tool because it focuses on how user and AI agents' applications behave rather than relying solely on predefined signatures, continuously monitoring AI interactions and identifying unusual patterns with the detection of potential risky activities, such as prompt injections or attempts, and abnormal agent actions.

It also detects unauthorized tool uses and data access behavior.

The governance and security aspects of Radware AI Security Suite help security teams by clarifying AI agents, tools, interactions, and dependencies, as well as identifying potential risk before a security incident, maintaining the inventory of AI assets, mapping relationships between agents and tools, monitoring usage patterns, and generating compliance-oriented reports.

In terms of accuracy and reliability, particularly in behavioral analytics, threat detection, and AI runtime protection, these are highly accurate capabilities of Radware AI Security Suite. Regarding reliability, it does not rely solely on static signatures but analyzes behavior, intent, and interaction across AI agents, tools, and users, improving detection quality and reducing unnecessary alerts.

Radware AI Security Suite is a strong tool, but it could be improved in the customization of the dashboard and reporting, which would help us tailor the views for different stakeholders and compliance requirements.

I have been using Radware AI Security Suite for one year.

Radware AI Security Suite is stable.

The scalability of Radware AI Security Suite is very good.

Customer support for Radware AI Security Suite is good and promptly supportive. I rate the customer support a ten out of ten.

I did not previously use a different solution.

The factors influencing our decision between API integration (out-of-path) and code proxy integration (in-path) for sensitive agent actions included security, deployment complexity, performance impact, and visibility.

We have seen a return on investment with Radware AI Security Suite, particularly in terms of time saved.

My experience with pricing, setup cost, and licensing for Radware AI Security Suite has been very good overall, with positive impressions regarding licensing and cost.

We evaluated other options before choosing Radware AI Security Suite, but details about those options are confidential.

I rate Radware AI Security Suite a nine out of ten. I rated it a nine because it provides strong visibility, governance, and real-time protection for AI environments while helping address emerging threats such as prompt injection, jailbreak attacks, and data leaks, and its behavioral analytics, AI agent discovery, and runtime protection capabilities give our security team greater confidence in managing Copilot and other AI agent deployments. My advice for others looking into using Radware AI Security Suite is to start with visibility and governance objectives before focusing solely on threat prevention, as adopting this solution will help with understanding which AI agent tools and integrations are open within the environment, making it very helpful for any kind of organization.