Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
donavanm
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
56 ms
·
601.
▲
by
donavanm
12y ago
Basically yes. Check out the kprobes docs for a nice description of how these frameworks work , https://www.kernel.org/doc/Documentation/kprobes.txt . Being able to intercept (and mangle) kernel function calls is aw
602.
▲
by
donavanm
12y ago
Check out exercism.io. Its effectively a series of unit tests where you fill in the implementation. Peers are there to help/comment on yiur specific implementation. As you work through the tasks they increase in scope and complexity.
603.
▲
by
donavanm
12y ago
Doing a public post mortem is nice. But there's a bit of, uh, "derp" in the post. 1) Conflating MTU and MSS. MTU is the maximum size of the layer 3 IP datagram. MSS is the maximum size of the layer 4 TCP data section. Specifi
604.
▲
by
donavanm
12y ago
Ok... So looking at IFG as 96 bits or 12 bytes of "overhead" thats 0.8% or 0.13% for 1.5k and 9k frames. Why do I care about 0.67% of throughput? And pretty much all silicon in over a decade does line rate at 1k anyways. Or if its
605.
▲
by
donavanm
12y ago
The atom server boards released last year actually have 4x 2.5g lanes. As far as ive seen everyone just uses 4x 1g serdes on them instead of the hybrid 10.
606.
▲
by
donavanm
12y ago
Pretty much everything has supported 9k jumbos for over a decade. The internets mostly 1.5k MTU, but you normally arent doing multi gigabit streams over public connectivity. The other argument is TSO. Your kernels probably writing a 64K &qu
607.
▲
by
donavanm
12y ago
Maybe SR/LR is the right choice for home. In the DC/rack some sort of DAC is the way to go. DAC will save you like 5 watts per port over optics.
608.
▲
by
donavanm
12y ago
> There also aren't many copper 10 gig top-of-rack switches, just the Cisco Nexus 3064T and Arista 7050T come to mind. Juniper doesn't even have one. I might be missing your definition of tor. The juniper QFX5100 series has the
609.
▲
by
donavanm
12y ago
The best questions have levels. Speaking to recursion specifically I had Flood Fill as a question the last time I interviewed. I solved it with recursion as it's the more ... "elegant" method. Then the interviewer asked about
610.
▲
by
donavanm
12y ago
+1, talking about rest endpoints hosted by junos is missing the forest for the trees. Protocols like Openflow, and whatever the contrail version was called, seems o be where network automation is headed. Centralized state & modelling a
611.
▲
by
donavanm
12y ago
With regards to the militarization of the Unites States executive branch I blame the Department of State. State has been an utter failure in their core competency of nation building for the past 30 years. Ethiopia and the Baltics in the &#x
612.
▲
by
donavanm
12y ago
Private armies aren't new. The very term "company" was originated as an organized, private, group of men under arms. I haven't read _The Modern Mercenary_, but it sounds a bit myopic. For example, calling out potential a
613.
▲
by
donavanm
12y ago
Project predated oracle, back in the osx 10.5-10.6 days when there was the dtrace integration as well. There were internal builds of the zfs kext and at one point I saw leaked source on the Internet. Project was abandoned because of legals
614.
▲
by
donavanm
12y ago
The magic homomorphic system doesn't exist, and I don't see any reason to believe it will exist on a time frame I care about. Short of that we can either deploy our own infrastructure (capital) or lease it from someone else (opex&
615.
▲
by
donavanm
12y ago
Speaking of, I think we'll see nvme being the new hotness this year. NVME is going to bring crazy parallelism and queue depth to block devices. Finally we'll be able to exploit flashs inherent parallel access in a sane way. http:
616.
▲
by
donavanm
12y ago
There are a ton of great napoleonic fortifications in the rhineland. A couple years ago I was touring through and stopped for an afternoon wandering around Neuf Brisach. Didnt realize it was one of his last projects.
617.
▲
by
donavanm
12y ago
Sure. By "trident" I really mean any merchant silicon switching platform. The Broadcom Trident ASIC/chipset really kicked this market segment off in 2011/2012ish. I mentioned it specifically as products like the Juniper
618.
▲
by
donavanm
12y ago
That's how most/all the congestion control works. It's essentially a send side only algorithm. On linux congestion control algorithm is actually a per socket option, as I recall. The receivers input/signal is limited to
619.
▲
by
donavanm
12y ago
Check out the Intel Atom Avoton and Rangley SOCs. Nice x86 cores, ECC, crypto acceleration, VT-x, passive TDP, and 4x 1/2.5gbe or 1x 10gbe depending on the serdes. I only wish they had VT-d to get sr iov. If you really need more connec
620.
▲
by
donavanm
12y ago
Most of those examples were tech demonstrators and reference designs, AFAIK. A small number of units, if any, were available. I don't recall many partners bringing them to market. Intel's not in the business of selling a couple th
621.
▲
by
donavanm
12y ago
And what would the action be when that limit is reached? Terminate all your EC2 instances? Delete your objects from S3/Glacier? Destroy all your RDS/DynamoDB/SDB tables? Disable your CloudFront CDN assets? Stop returning DNS
622.
▲
by
donavanm
12y ago
I havent read the paper yet, but Im a little surprised by the hyperloglogs. I was under the impression that HLLs break down when your symbol frequency varies by orders of magnitude. Those are exactly the patterns id expect to see in block&#
623.
▲
by
donavanm
12y ago
Sorry, I dont think ive seen that. But we have offices all over and do sponser where appropriate.
624.
▲
by
donavanm
12y ago
We have people working on everything from firmware, network devices, kernel and virtualization development, up to applications doing millions of tps and serving billions of bits per second. Sounds like it's too late for internship but
625.
▲
by
donavanm
12y ago
I'm curious what you're doing that's substantially cheaper with Google Cloud DNS. The base pricing looks the same, except Google being $0.2/month and AWS $0.25/month per hosted zone. Also keep in mind that queries f
626.
▲
by
donavanm
12y ago
The dst/route cache has effectively been replaced by the TCP metrics cache. Values like MSS, SRTT, and cwnd are still useful to preserve across streams between peers. See ip-tcp_metrics(8) for a decent summary. David Miller has a summa
627.
▲
by
donavanm
12y ago
cloudFront has been cheaper than S3 for as long as I can recall. And it makes sense with the way amazon does cost based pricing. Its cheaper to transfer cached bits fromthe edge of the network than pull it from inside of multiple datacenter
628.
▲
by
donavanm
12y ago
Id hazard the most prolific use case of md5 as http checksum is AWS S3. You can, IIRC, send content-md5 on PUT and POST and theyll reject with a 400 if the body hash doesnt match. Conversely the default etag is the body md5.
629.
▲
by
donavanm
12y ago
Color me confused. Just a week ago they posted "We also have server-side means to “hash away” and route around trouble spots, if they occur." https://code.facebook.com/posts/360346274145943/introducing-..
630.
▲
by
donavanm
12y ago
Broken TCP window scaling? The connection should still work, with the sender backing off right? If you mean busted PMTUD yeah thats awesome when the handshake and GET works, but you cant get data back.
More ›