Today’s Security Requires Specialized Processors

Many are familiar with central processing units (CPUs), graphical processing units (GPUs), and maybe even data processing units (DPUs). But what is a security processing unit (SPU)? As the name suggests, it's a specially designed processor focused on making security tasks run faster and more efficiently than relying only on a CPU alone. An SPU can act as a self-contained system to better meet the requirements of smaller environments such as a branch office.

The main advantages of security systems designed with SPUs versus CPUs are:

1. Superior price-performance
2. Ability to run applications simultaneously
3. Increased power-efficiency
4. Smaller footprint
5. Built-in security

Let’s look at some of the processors we already rely on so our devices can keep up with the escalating demands of today’s applications and services.

General Purpose Computing

The overall design of generalized computers has not really changed for a few decades. CPUs, memory, buses, and peripheral cards usually perform specific functions, such as supporting network interfaces or providing graphical processing.

Central Processing Units

A CPU is the primary component that performs most of the processing inside a computer. It carries out computer program instructions by performing basic arithmetic, logic, control, and input/output operations specified by the instructions. The CPU is often referred to as the “brain” of the computer. But while it is very good at generalized tasks, it cannot match the performance and efficiency of application-specific integrated circuits (ASICs) for specialized tasks.

Graphical Processing Unit

A GPU is a specialized electronic circuit designed to rapidly manipulate and alter memory to accelerate the creation of images in a frame buffer intended for output to a display device. GPUs are used in embedded systems, mobile phones, personal computers, workstations, and game consoles and are typically efficient at handling many repetitive tasks simultaneously. This makes them well-suited for tasks like image processing, 3D rendering, and machine learning computations. More recently, they are being used to process large language models (LLMs) for AI.                       

Data Processing Unit

A DPU is a specialized hardware component designed to accelerate data processing tasks in computing systems. DPUs are typically used in data centers to offload and accelerate specific workloads, such as networking, storage, or security functions. By handling tasks separately from the main CPU, they can improve performance, reduce latency, and increase efficiency. SmartNICs are used in a similar way for offloading purposes to achieve accelerated network processing.

Specialized Security Computing

These general-purpose computing elements have dramatically increased in speed over the past decade, however they still lack the ability to support many basic security processing functions. While different software components may be able to help the CPU, specialized security-focused tasks are still something that cannot be accelerated well by these commercial off-the-shelf components. Fortinet is the first vendor to develop custom ASICs explicitly designed to support critical security functions. We have been developing specialized versions of these ASICs for 20+ years, and most importantly, we have developed the software and operating system (FortiOS) needed to make them all work together at high speed.

Let’s take a look at the portfolio of specialized processors we have developed:

To benchmark the security processing performance, we gather data sheet specifications from different vendors to calculate a general Security Compute Rating for each function. The latest FortiGate model to be released with the SP5 is the 200G. It leverages the SP5 in two ways: as an NP7 Lite and as a CP10.

Fortinet SPUs outpace other solutions with the highest Security Compute Ratings for benchmarks such as:

• Stateful firewall throughput
• IPsec VPN performance
• Concurrent sessions
• Sessions per second
• Inspection of encrypted data
• Power-efficiency

Like all other SP5-based FortiGate models, the 200G can accelerate 14 different applications, which would be impossible using a standard CPU.

This table compares the top firewalls on the market against the target performance numbers of the FortiGate 200G series, which leverages the SP5.

Additionally, the power efficiency of the FortiGate 200G is 4x the industry average. With some customers deploying hundreds, even thousands, of these devices, those power savings really add up. For equivalent performance, many competitive models in this price range are 2U rather than 1U boxes, which can use up rack space and require additional cooling to keep them running efficiently.

Summary

Although not as well-known as the GPU, the SPU has a similar objective: offloading the CPU or, in some cases, becoming a SoC to completely replace the CPU. At the end of the day, the role of GPUs and SPUs is to accelerate critical functions to keep up with application and environmental demands. When it comes to security and the rate at which a cyber incident can occur, speed is of the essence. Traditional security devices that rely on general-purpose CPUs simply cannot keep up, meaning you pay significantly more for equivalent performance.

With network performance and security demands continuing to grow and the rate of cyberthreats escalating, it is essential that your security solutions are able to stay ahead of your cyber adversaries.

Set up a call with your Fortinet representative for a demonstration of the power and performance of the industry’s only SPU-enhanced security solutions.