Recently, scientists and engineers started exploring possibility of using GPUs (Graphics Processing Units) in massive data processing other than graphics workloads. GPUs are heavily used for HPC (high performance computing) or many computation-intensive applications.
Some of the early works also find that GPU is useful for accelerating cryptographic computation such as AES, SHA1, and RSA. Although GPU's high computation capacity delivers higher throughput compared to that of CPU, accelerating SSL poses further challenges due to interactive nature of network applications. GPU's high computation power comes from massive number of small cores, and each core is not as fast as CPU cores. Thus, fully exploitation of the GPU's power requires parallel processing of many independent operations and may end up increasing response time, which is not acceptable in the interactive network applications.
"The key idea is to send all requests to CPU when the number of pending cryptographic operations is small enough to be handled by CPU," the research team says in an earlier paper. "If requests begin to pile up in the queue, then the algorithm offloads cryptographic operations to GPUs and benefits from parallel execution for high throughput."
SSL transactions per second (TPS) using just the CPU on the test servers totaled 3,632 in one experiment. Using the proxy GPU and their algorithm yielded 18,482 TPS. The group used an Intel Xeon X5550 CPU with four cores and an NVIDIA GTX 480 graphic card with 480 cores.
We are already planning and working on about making our SSL offloading with GPU available...
TRscaler born in Turkey, offers industry-leading solutions for Information Technologies and efficiently with no unexpected costs.
Subscribe to our newsletter and stay up to date with the latest news and deals!