[cs615asa] Some Thoughts about New York SRE Tech Talks

[Yangmu Jiang]

Hi all,

I’m here to share some of my thoughts and feelings about the New York SRE Tech Talks:


Why I chose it?

	Actually I didn’t choose anything. There was no alternative for me, I just followed professor’s suggestion…


What I learned?

	On the first talk, I sat upright. I heard words like eBGP, Layer 3, migration, failure tolerance, implementation and configuration management.

	But, I can’t say I understand any of them.

	I don’t know what eBGP is, what we should keep in mind when talking about failure tolerance, configuration management, etc. I prefer not to put so many high level concepts in mind when I don’t have any related experiences.

	—

	The second talk was about kernel testing. Since I had some experiences on low-level development, this topic aroused my interest. Not so surprise, the speaker talked about why/what/how on kernel testing.

	Again, knowing our goal first! Performance! Stability! Security! Quantifying everything, this is what we call engineering. It simplifies things, but also make new learners feel boring.

	We can’t do much thing without a good tool, it’s still true in kernel testing. The speaker introduced many different types of testing tools. What interests me most is ‘flexible IO tester’ and ‘fuzzer’, I’m not very familiar with these two things. My current understanding about ‘fuzzer’ is that it’s a test program that generates tons of invalid input to a target program in order to test its robustness. Obviously, the more things we can monitor and log, the easier test will be.

	On the other side, there are still tons of tedious work on kernel testing. It seems this is why the community build forums.

	—

	The third talk, all I got is: it’s a bad idea to make many copies of auth keys. What’s bootstrap lock? What’s data dependency lock? I don’t know. Or in other words, I’m not interested in this topic for now…


Final Words = This is a very subjective summary. It feels good! :)


Sincerely,
Yangmu Jiang