Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Joyent Public Cloud EOL

docs.joyent.com

Post reply on HN
Print view View original

251–260 of 289 posts

Re: Joyent Public Cloud EOL

#251

Post by cbsmith »

Annatar wrote: post #233
cbsmith wrote: post #224

Earlier quoted context omitted.

I'm sorry, but you don't know what you are talking about. JavaSpaces was a distributed key-value store. JXTA was all about a resilient distributed mesh. I hate on NFS all the time, but WebNFS demonstrated a much better understanding of how to do reliable data delivery over the Internet than HTTP ever did. Heck, even the old stuff around NIS+/ONC or NeWS or USENET covers distributed systems pretty well. If you compare…

"If you compare Sun to their competitors at the time: IBM, DEC, SGI, HP, and even Microsoft. Sun was way ahead on distributed systems." Careful now, you're on thin ice: Sun Microsystems might have been a pioneer in distributed cloud computing, but nobody did it better than SGI: once they had the NUMALink technology from CRAY, they invented cluster routing in hardware which would linearly scale with added router modul…

That's not the distributed computing that the OP was referring to. Explicit reference was made to WAN-style networking that makes the crossbar switch effectively seem like motherboard traces.

Re: Joyent Public Cloud EOL

#252

Post by Annatar »

zzzcpan wrote: post #220
KirinDave wrote: post #217

Earlier quoted context omitted.

Why does ZFS reflect a lack of experience in the internet and/or distributed systems? Simply because it didn't follow a questionable tech trend of "everything is a DHT?"

Because they didn't seem to know any of the simple and nice ideas from distributed systems to solve pain points of managing space, since adding, removing nodes, sharding/partitioning is something you have to figure out pretty much from the start.

Distributed systems and write arbitration on the same file are still terra incognita. This has never been solved because no one, and I really mean literally no one has been able to solve write arbitration to the same file or even the same block from multiple sources at the same time yet. Especially not in a generic way. Which node's update is the correct one in that instance? There is an entire world of computer science research waiting there. I'm sorry to have to write this, I really am, but you have no idea what the problems of the systems you are referring to are. It would be enough for me to cut off electricity to one or even all of your systems and then you would see all hell break loose on your "simple and nice distributed system ideas".

You are just blissfully unaware of all the issues because you assume the network and the hardware are infallible and apparently you gave no thought to write arbitration in a true clustering filesystem. You have no idea what lurks beneath the surface and how fragile your "simple and nice distributed ideas" are, especially when implemented in software.

Re: Joyent Public Cloud EOL

#253

Post by Annatar »

cbsmith wrote: post #251
Annatar wrote: post #233

Earlier quoted context omitted.

"If you compare Sun to their competitors at the time: IBM, DEC, SGI, HP, and even Microsoft. Sun was way ahead on distributed systems." Careful now, you're on thin ice: Sun Microsystems might have been a pioneer in distributed cloud computing, but nobody did it better than SGI: once they had the NUMALink technology from CRAY, they invented cluster routing in hardware which would linearly scale with added router modul…

That's not the distributed computing that the OP was referring to. Explicit reference was made to WAN-style networking that makes the crossbar switch effectively seem like motherboard traces.

Doesn't matter (and I don't care) what they referred to, only the end result counts: SGI designed and implemented cloud computing in hardware. Color me impressed. Haven't seen anyone successfully do it since, hardware- or software-wise (and everyone is sure as hell trying!)

Re: Joyent Public Cloud EOL

#254

Post by Annatar »

cyphar wrote: post #250
Annatar wrote: post #248

Earlier quoted context omitted.

Sorry buddy, but https://illumos.org/man/1/filesync predates rsync. Borgbackup is a third party, unbundled application. I don't see what it has to do with people hacking on GNU/Linux like mad. "I just think it's quite ludicrous that you're saying that any engineer who didn't work at Sun pre-2010 never invented anything." You didn't get it quite right: I'm saying that professional engineers worked at Sun Microsystems…

Dave Chinner is an ex-SGI engineer who has been working on XFS since the beginning of the project and is currently a Linux kernel maintainer. Brendan Gregg is ex-Sun and worked on DTrace and currently is helping improve Linux's tracing capabilities. Are they "amateurs who are incapable of learning" or "professional engineers"? I've seen comments from you about how GNU/Linux is awful in every respect ever since you cr…

Dave and Brendan are former SGI and Sun engineers. They are refugees. They gave up. So once again we're back to what I said: either one came from SGI, or hp, or Sun Microsystems, but one sure as hell didn't come from the GNU/Linux being an engineer, nor will such a thing ever happen. For 20 years they've been hacking furiously on Linux and still the basic things don't work correctly. That's what happens when one hacks instead of system engineering solutions.

"I really can't imagine being so stuck in a mindset that you feel the need to spend so much of your time being angry about such a large community of people."

Computers were my life, my passion, my calling. They destroyed all that passion in me with their GNU and their GNU/Linux. I will not forget. I will not forgive. They made my professional life a living hell. I hate working in IT because of those people and their mentality.

"Honestly, I just don't get it and I earnestly hope that this is just an online persona you have."

  You must be the change in the world you wish to see.
And I believe that, deeply. I'm absolutely convicted about it. Even if I single handedly must teach apprentices, I will continue to fight until the last drop of blood against GNU and against GNU/Linux and teach people, person by person if I must, as I have been, showing them one on one how awful it is and how superior and easy a real UNIX like SmartOS is. Either until I die or manage to get out of IT. Because truly awful things like GNU and GNU/Linux deserve to be relegated to dark history of IT, and some good and beautiful, well thought out things like SmartOS are worth fighting for. I will continue to fight against ignorance and trend pandering and for enlightment. I will strain like buddha strained if I must. Some things are just worth fighting for and some things one must make a stand against, no matter the cost to oneself. I am forced to work on and with GNU/Linux because people think like you, but I will fight it and resist it every step of the way and I will continue to teach because working on Linux and with GNU/Linux is more work, not less work for me. That alone is more than enough to motivate me.

Re: Joyent Public Cloud EOL

#255

Post by zzzcpan »

Annatar wrote: post #253
cbsmith wrote: post #251

Earlier quoted context omitted.

That's not the distributed computing that the OP was referring to. Explicit reference was made to WAN-style networking that makes the crossbar switch effectively seem like motherboard traces.

Doesn't matter (and I don't care) what they referred to, only the end result counts: SGI designed and implemented cloud computing in hardware. Color me impressed. Haven't seen anyone successfully do it since, hardware- or software-wise (and everyone is sure as hell trying!)

Cloud computing in hardware is oxymoron.

Re: Joyent Public Cloud EOL

#256

Post by zzzcpan »

Annatar wrote: post #252
zzzcpan wrote: post #220

Earlier quoted context omitted.

Because they didn't seem to know any of the simple and nice ideas from distributed systems to solve pain points of managing space, since adding, removing nodes, sharding/partitioning is something you have to figure out pretty much from the start.

Distributed systems and write arbitration on the same file are still terra incognita. This has never been solved because no one, and I really mean literally no one has been able to solve write arbitration to the same file or even the same block from multiple sources at the same time yet. Especially not in a generic way. Which node's update is the correct one in that instance? There is an entire world of computer scie…

Flagged your comment for this. You should at least glance over my comment history on distributed computing before attacking me on things you don't understand.

Re: Joyent Public Cloud EOL

#257

Post by aramallo »

zzzcpan wrote: post #219
timClicks wrote: post #213

Earlier quoted context omitted.

"The network is the computer" doesn't feel like a slogan that a company ignoring distributed systems would use

I guess I rephrase: they didn't do any of the distributed systems stuff necessary for the cloud if people so desperately want to call all the network stuff they did "distributed systems". Technically they were distributed systems, sure, just like anything client-server, but irrelevant to the challenges of fault tolerant geographically distributed asynchronously communicating systems.

I completely disagree, see my comment above. I was doing asynchronous distributed data replication and compute over a grid in 2001 using Jini and Javaspaces, and I was the one late to the party.

Re: Joyent Public Cloud EOL

#258

Post by zzzcpan »

aramallo wrote: post #257
zzzcpan wrote: post #219

Earlier quoted context omitted.

I guess I rephrase: they didn't do any of the distributed systems stuff necessary for the cloud if people so desperately want to call all the network stuff they did "distributed systems". Technically they were distributed systems, sure, just like anything client-server, but irrelevant to the challenges of fault tolerant geographically distributed asynchronously communicating systems.

I completely disagree, see my comment above. I was doing asynchronous distributed data replication and compute over a grid in 2001 using Jini and Javaspaces, and I was the one late to the party.

So old broken networking tech (Jini and Javaspaces) over a fast reliable local network (a grid) in 2001. How could this knowledge help anyone with fault tolerance, high-availability, partitioning/sharding, CAP consistency/availability trade offs, etc., which internet companies at the time were figuring out?

Re: Joyent Public Cloud EOL

#259

Post by Thaxll »

Annatar wrote: post #249
Thaxll wrote: post #198

Earlier quoted context omitted.

I see multiple problems with that way of thinking, first HW will fail and AWS, Google ect ... showed us that you should not invest in very sophisticated things for those kind of problems. It feels that you're building something to avoid HW issue but they will happen no mater what so it's better to invest in HA with a mix of standard servers + SW. And solaris zones... all the modern tooling are built around Linux / Do…

Because with Triton or just imgadm and vmadm one doesn't need Kubernetes. Solaris zones, imgadm and vmadm make Docker and Kubernetes obsolete. "if you think with a small company you can change that you are very wrong" You must be the change you wish to see in the world. Mahatma Gandhi

You don't know what you're talking about if you think those are equivalent. And well their public cloud is pretty much dead so good innovation!

Re: Joyent Public Cloud EOL

#260

Post by notacoward »

Annatar wrote: post #181
zzzcpan wrote: post #127

Earlier quoted context omitted.

> Technically speaking, they were quite far ahead of the curve. While VMs were all the rage, they were shipping container-based products > The technical abilities of the Joyent folks were well beyond just about anybody else for a long time. None of that was ever true. Whether they had container-based or VM-based tech was a tiny unimportant detail for a cloud company. Although over engineering in this area certainly m…

Sun Microsystems literally invented the network is the computer . They were the first company ever to even begin to think in terms of distributed systems on a network scale (DEC had done some pioneering work on the hardware front prior to that with VMS, but not the network). The network file system, NFS, exists because Sun Microsystems engineers willed it so -- they wanted to multiply the computational power of small…

Wow, they invented the slogan. Sun alumni seem awfully thin-skinned about not getting credit for every little thing they did, which is why it's pretty unforgivable to deny DEC credit for anything more than "some work on hardware". I was working on DECnet-based remote file access software in 1989. It wasn't new, but NFSv2 was. DEC already had real clusters doing real work before Sun did. Within the company there was already a globe-spanning LAT network. I had friends at Apollo who were also way ahead of Sun wrt distributed computing. AT&T's RFS was at least contemporary with NFS, as were Netware and Vines. Sun was definitely a player in that space and made some seminal contributions, but to say they were the first even to think about it is unmitigated BS.
Post reply on HN