9/24/2023 0 Comments Spectre performance updatesTest the prepared disks (filled with random data using DD utility): Intel DC S3700 SERIES and Samsung SSD 960 EVO NVMe M.2 using FIO and DiskSpd tools.Measure CPU and RAM performance with AID64 and WinSat.Check the system for vulnerabilities prior to testing with InSpectre.To make my intentions clear for everyone, here’s what I’m gonna do step-by-step: Tests will be run on Dell PowerEdge R720 with Windows Server 2016 operating system. And finally, the InSpectre 6b tool will be used to check if the vulnerabilities are present in the system. These will also allow us to investigate the RAM performance in detail. Further, to create a RAM disk, I’ll go for DataRAMDisk 4.4.0.36 from Dataram inc and StarWind RAM Disk 5.8 from StarWind Software. I’ll also use DD (Disk Dump) 0.5 tool to fill the tested disks with some random data to increase the benchmark accuracy. To compare the RAM and HDD parameters, I’ll use FIO 3.4 from Jens Axboe and DiskSpd 2.0.17 from Microsoft. I’ll use these to control CPU and RAM parameters. The testing result will be based on the following criteria:Īccording to my experience and taste, well, sorry for that but that’s how it is, I’ve decided on four utilities: Aida64 from Final Wire Ltd and Windows Server 2016 embedded utility – Winsat. That’s why I’ve decided to run a series of tests on Dell PowerEdge R720 server before and after patches installation to measure their performance influence. For a system administrator, performance may be just as important as security. When searching for protection against these vulnerabilities, I’ve hit several articles claiming that Microsoft patches may cause to downgrade your hardware performance. Instead, he calculates the mistakenly cached bytes in his address space, checking read latencies in different lines of the previously created volume. The adversary doesn’t read the data itself (he can’t read cache from a random location). Using different tricks, the processor caches byte values from the memory address unavailable for the attacker. Potentially, hardware vulnerabilities can affect the performance of the hardware which leads to possible issues with the CPU-intensive workloads.Īnyway, the most important thing is that under all scenarios, the aim of the attacks is cache. These were formally described as similar vulnerabilities: Meltdown (first pattern) and Spectre (second and third). The blog pointed at three attack patterns on the CPU cache: Meltdown and Spectre attacks were noticed by Google Project Zero analysts. The bad news is that it potentially allows local applications (when executing a special program) to get access to the content of the virtual memory of current application or other programs. So, Spectre is a hardware vulnerability that affects most of the modern microprocessors, namely х86/x86_64 Intel and AMD and certain ARM CPU. The vulnerability doesn’t affect the microprocessors from AMD and has an official reference CVE-2017-5754. It allows an adversary (when executing a special program) to get unauthorized access to read from the memory used by the OS kernel. Meltdown is a hardware vulnerability found in the number of processors, namely in Intel and ARM family processors. Before we take a deep dive into the tests and numbers, let me tell a few words about Meltdown and Spectre and outline the testing scope to make sure we speak one language. But not about the harm they cause, this has been covered widely in numerous articles, but how Microsoft patches intended to protect you from the vulnerabilities, affect (if they do) the hardware performance. In today’s topic, I’d like to talk about the Meltdown and Spectre vulnerabilities.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |