As you might know by now, Meltdown and Spectre are two significant hardware bugs that enable hackers to exploit critical vulnerabilities in modern processors. They have been reported by various, independent security teams and extensively relayed in the press. The scope of the issue is very important as there are very few unaffected computers in the world.
But while consumers should definitely be aware that even Mac systems and iOS devices are at risk, the news has been particularly concerning for entities that manage the data of others. At the end of the day, the bugs enable programs to easily steal the data that is processed on affected computers, including your AWS, Azure or GCP instances.
Since they’re the product of years of hardware engineering decisions, it will take as many years for a fundamental change in how we architect our processors to give birth to a new generation of processors. Furthermore, the idea of potentially having to replace hardware components at such scale is arguably one of the most daunting procurement operations ever considered in the history of IT.
For now, and as a consequence, software patches will provide an initial protection against the important risks introduced by these vulnerabilities…
In exchange for up to a 30% degradation in performance of your microservices.
Quantifying the Meltdown and Spectre Performance Degradation
At Stacktical, we’re developing a platform that models the performance of your microservices using predictive technologies. Predictions make everything faster and easier. Our customers can run more tests than they could initially afford to, and automate their interpretation into the metrics that matter to them. For example, we’ve recently helped a cryptocurrencies company sustain 5 to 10 times more signups per second with their API.
Following up on Meltdown and Spectre impacts on server performance, we’ve started working on a scalability experiment. The objective is to quantify the decrease in concurrent users, transaction per seconds and response time experienced by monoliths and distributed, microservices based architectures as they’re getting patched.
After we’ve gathered a sufficient amount of predictive data, we will share our conclusions with the community through a set of Stacktical Scalability Reports, similar to this one.
We’d love to have some devs, ops and companies participate in the experiment. If you’re interested, don’t hesitate reach to our founding team at [email protected] and spread the news in your network.
Our team thanks you all for your help!