HNHacker News
TopNewBestAskShowJobs

nhoughto

451 karma · joined October 9, 2014

submissionscomments
nhoughto··on We clone a running VM in 2 seconds
great to hear more details about snapshot/restore in the wild, plenty about firecracker but seemingly much less about this exciting feature in real usecases.

looking forward to the unwritten details / future posts too, particularly:

- How to handle network and IP duplicates on cloned VMs

and

- Turning a Dockerfile into a rootfs for the MicroVM (quickly)

nhoughto··on What I learned working with a senior engineer as a new grad
Maybe I'm the outlier but at >15 years into this job I think back on what I was like/did/thought 6 years in and shudder, at that point I never mentored anyone my junior seriously enough to influence their perception (I don't think) but if I did I'm not sure it would have been a positive influence. The lack of failures, felt pain, time to reflect.. in combination with the hubris of 20-smth youth isn't a recipe for good software engineering I don't think.
nhoughto··on LocalStack 1.0
Yeah we gave up when we got to using lambda. Even if the api works it can’t scale the same way!
nhoughto··on Ask HN: How to get developers and UI designers to work well together
Sympathy is what it all boils down to.
nhoughto··on Apple previews Lockdown Mode
ironically if this is successful we likely won't know how deep the hardening goes, unless security researchers who have legit access release something?

but i'd love to know the series of changes that goes into something like this if its not just hype, potentially unpicking a giant complex system from the bottom up to remove attack surface.. interesting challenge!

nhoughto··on Code replay attack on the myGovID Scheme
Worked on something very similar and we chose not to move ahead with this kind of feature because of the phishing risk, without redefining the problem (like webauthn, Fido etc) it is very easy to dupe a user / replay attack. Most users don’t pay that much attention and friction hurts adoption.

Kind of funny the govt shipped the feature even with the spam risk and our more commercially oriented competitor product didn’t. You’d Hope they would be raising the bar for security not lowering it.

nhoughto··on 10 Years of Meteor
I built the exact same thing.. and had exactly the same outcome, crazy coincidence! Probably my biggest failure and learning experience so far.
nhoughto··on Why are nuclear power construction costs so high?
And also they already can, if you believe it:

https://www.bloomberg.com/news/articles/2019-05-15/disney-wo...

nhoughto··on Connect: A Better gRPC
Good stat, anymore detail? Love to read more into why
nhoughto··on Internet drama in Canada
Australias NBN for all its painful mistakes at least had the structure right. National wholesale network with strong regulatory power to drive efficiency. Lots of problems but consistent, national problems that require solutions because hurt many people. Canadas problems sound different.
nhoughto··on AlloyDB for PostgreSQL under the hood: Columnar engine
Interesting they touched on a couple of pain points I’ve had with spark execution, not quite the same thing but I’ve wondered why some of the way it execute isn’t smarter. Bloom filters for cheap way to filter big joins, using single scans to do multiple jobs in the plan.. I’ll have to have a look.
nhoughto··on Speeding up sort performance in Postgres 15
Interesting to see there is still some more to squeeze out of the old girl yet. Most seem like things that could have attempted before now but I guess weren’t the top pain points to fix, I wonder how many more 4-6% savings there are in there, good to see people going deep and keeping the improvements coming.
nhoughto··on AWS Lambda function URLs: Built-in HTTPS endpoints
So lambda invoke direct via api has 6mb limit: https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-...

But api gw is 10mb, not for lambda tho, confuse: https://docs.aws.amazon.com/apigateway/latest/developerguide...

Use s3 object lambda and have no limit to push to s3 with logic, adds a bit of latency tho https://aws.amazon.com/blogs/aws/introducing-amazon-s3-objec...

All historical but hard to understand the limits as a poor user..

Ended up going with cloudflare workers for s3 uploads with extra logic just to avoid the unknown, workers are great btw.

nhoughto··on AWS Lambda function URLs: Built-in HTTPS endpoints
Yeah didn’t see anything about timeouts or body size limits.. is a good question. Lambda + apigw vs invoke vs alb vs the rest having different limits has got me a few times.
nhoughto··on AWS Lambda function URLs: Built-in HTTPS endpoints
Yeah you need an aws iam identity to call lambda invoke, for public consumer this is better.
nhoughto··on `COPY –chmod` reduced the size of my container image by 35%
Works for lots of examples where you are packaging a statically-ish linked thing into a container, could be node/golang/python/java etc. Def lots of other scenarios where it doesn’t work, but sometimes you can push that to a base image built differently. Keep the majority of things simple, less footguns.

(Edit: I realize most of those aren’t statically linked, better description might be “things copied straight into container, not installed”)

nhoughto··on `COPY –chmod` reduced the size of my container image by 35%
Do it all in Userland with jib cli, doesn’t work for all cases but it’s constraints generally keep you honest.
nhoughto··on Nintendo Switch prevents downgrades by irreparably blowing its own fuses (2020)
Ah I see, thanks for the insight. Didn’t realize they were like a storage primitive like that (obvious now..)
nhoughto··on Nintendo Switch prevents downgrades by irreparably blowing its own fuses (2020)
ah right so they do this to avoid having a proper secure enclave, like a very focused secure storage capability related to what the efuses are logically related to. Makes sense, I guess i just assumed they would have a secure enclave like phones etc as consoles are one of the original 'trusted computing' devices that people buy and obviously to avoid piracy etc having it work properly is important but also hardware BOM is a consideration too.

I guess the secure enclave having storage introduces another attack too, wiping/corrupting/replacing that storage somehow, thus efuses, simpler and more straightforward.

nhoughto··on Nintendo Switch prevents downgrades by irreparably blowing its own fuses (2020)
Why does it need to physically modify the hardware via melting fuse when that fuse is read by enclave / boot loader code itself? If trusted code is trusted then couldn’t it store its state securely without melting fuses?

I must be missing something, either the bootloader execution is trusted and should be able to store state securely, Secure Enclave style, or it’s not and melting things doesn’t solve the problem as compromise of the code means the fuses can be ignored..

nhoughto··on Ask HN: Who operates at scale without containers?
seemed more to be technical and organisational. More and more teams all the time, more and more products, more and more technical requirements / tuning etc, needed a strong abstraction to reason about it all.

Homelab would be fine =)

nhoughto··on Show HN: Appwrite – Open-Source and Self Hosted Firebase Alternative
I had the same sentiment, I was thinking is this the localstack of the Firebase world? have you doomed yourself to replicate their problems? but no its a firebase _alternative_ not a replacement, well played. Lots of good explainer content too.
nhoughto··on Threads, Threads, and More Threads
Hmm feels like there is a bit of a misunderstanding on the meaning and cost of threads and how the OS schedules work from threads onto logical CPUs and what state they are in. It is expected that processes might use more than 1 thread, and it is expected that you would have more processes than logical CPUs. Can not-dead process exist without a thread at all? I'd expect not.

7000+ threads does seem like a lot though and agree there is the general lack of caring about resource consumption in all forms, eg Electron apps. At this point having lots of threads seems almost expected by various OSs and something they solve for to reduce the cost of it so the user doesn't have to worry about it.

nhoughto··on My guiding principles after 20 years of programming (2020)
yep agreed.

I've often thought about what interesting findings you would get back if time and people were infinite and all those little things you thought, 'huh? thats weird' and put it in a queue somewhere and a few days later it had been thoroughly investigated and findings documented by some smart person. I'd expect you'd find some weird networking problems for why ssh hangs sometimes, misconfigurations, waste, security vulns etc. So many things we skip over day to day bc it would be a distraction.

nhoughto··on The Code Review Pyramid
I'll use this I think, very handy. I'd probably emphasis all the text/descriptions more too, lots of important nuggets in there that aren't obvious when glancing, like automating the top portions of the pyramid, humans shouldn't be mis-formatting code and others human comment on it anymore.
nhoughto··on Ask HN: Who operates at scale without containers?
Recent convo with Cloudflare they said they are mostly container-less, debian packages installed and running on the the host, which is impressive at their scale and complexity. They did seem to think that this approach was hitting its limits and what the future might look like when I spoke to ppl there recently (end 2021), assumption being the future could be containers.
nhoughto··on Problems with “graceful shutdown” in Kubernetes (2019)
ah good tip, I don't care about session affinity or custom balancing algos so that works. I'd imagine running in GKE or AWS you would also avoid the DNAT / conntrack overhead as pods by default use a routable VPC IP instead of a magic CNI IP. Would have to test that though.

Quote from related issue:

The NGINX ingress controller does not uses Services to route traffic to the pods. Instead it uses the Endpoints API in order to bypass kube-proxy to allow NGINX features like session affinity and custom load balancing algorithms. It also removes some overhead, such as conntrack entries for iptables DNAT.

nhoughto··on Problems with “graceful shutdown” in Kubernetes (2019)
Is this true of the OOTB GKE nginx ingress? Hard to tell by 'load balancer' do they mean nginx ingress reverse proxy?

I can imagine the delay between updating the GCP global load balancer service from GKE would be much higher than nginx-ingress reacting to changes in pod health/endpoints.

Either way I guess the takeaway is there is a race there between endpoints being updated and those updates propagating, and seems like that isn't handled as perfectly as people assume, and this likely gets worse with node contention and Kube API performance problems.

nhoughto··on The Windows Shutdown crapfest (2006)
I mean a monorepo is basically the modern answer to this, use technology to the solve the coordination problem, its just a different (better) answer. At Windows scale (like Google and FB etc) it would require effort but is do-able, and has been done elsewhere.

A multi tiered tree of repositories sounds very complex though, and would require its own effort, amazing it worked at all.

nhoughto··on Why aren’t there more dogs at the doctor’s office?
interesting to think about the amount of 'information' that is present in an environment (like smells) that we aren't privy too and don't understand. Its easy to assume it doesn't exist and then be surprised when a trained dog knows someone is sick immediately.

The information (in the form of chemicals etc) is there we just can't sense it via machine or otherwise, puts the idea of technological superiority into perspective.

← PreviousPage 2 of 9Next →