The Internet Is 'Built Wrong' 452
An anonymous reader writes "API Lead at Twitter, Alex Payne, writes today that the Internet was 'built wrong,' and continues to be accepted as an inferior system, due to a software engineering philosophy called Worse Is Better. 'We now know, for example, that IPv4 won't scale to the projected size of the future Internet. We know too that near-universal deployment of technologies with inadequate security and trust models, like SMTP, can mean millions if not billions lost to electronic crime, defensive measures, and reduced productivity,' says Payne, who calls for a 'content-centric approach to networking.' Payne doesn't mention, however, that his own system, Twitter, was built wrong and is consistently down."
The Internet Is 'Built Wrong' (Score:5, Interesting)
So was a 1932 Ford. So were the highways in 1932. So was an analog computer in 1959.
The only thing wrong about the internet is that it has become obsessed with money rather than information. Technical issues will be worked out over time.
Many Organically Grown Systems Are Non-Optimal (Score:3, Interesting)
With something as complex and "live" as the Internet, process is more important than paradigm: the real question is how to optimize from the current live state, rather than mumbling pointlessly about how it should've had better roots.
Shoulda but didna. So, let's move on.
Also, I tried to send this guy a tweet but all I got was a message saying, "I'm sorry, a problem has occured; please reload the page."
Wanker.
Re:"Content centric"? (Score:3, Interesting)
I got the impression that he was talking about divorcing the content from the presentation, which sounds fine in theory but a lot of people want to have more control of the presentation...That was kinda the point of HTML in the first place; we'd have stuck with Gopher if all we wanted was pure content with a static presentation.
Even in a modern context, we could have switched to XML to divorce the information from the presentation, and there hasn't really been a charge in that direction.
It's hard to say what he really meant because the whole thing is lacking in specifics.
Cute! (Score:1, Interesting)
Isn't that cute, programmers pretending they are real engineers, by doing little studies on large systems, while ignoring the failings of their own smaller programs.
Satisficing (Score:3, Interesting)
IPv6 is an interesting case study. Theoretically better, but largely unadopted. The net benefits cannot be large.
Too many projects have been killed by over-optimizing. And people who say something is impossible should get out of the way of those actually doing it!
n00b (Score:4, Interesting)
SMTP is a terrible example. Ultimately the users are the ones opening e-mails, getting browser-jacked and their passwords stolen because they don't know what is in front of them. Sure clients were the problem for a while but that "phase" has passed, developers have learned how to mitigate most attacks.
The only thing that is "wrong" fundamentally with the internet is the separation of DNS and the routing protocol.
For all intents and purposes a DNS failure causes a network outage. It also dramatically increases client latency when it is not configured correctly which look like network issues, but are not.
I'm sure when IPv4 was created the notion of mixing both services was unthinkable due to the additional amount of data needed to move names around at layer 2/3. This is no longer the case and we should really try to move away from a central naming system.
Re:How Is This Different From a CDN? (Score:3, Interesting)
I may be misinformed but how is this any different than a Content Delivery Network (CDN) [wikipedia.org]? I believe these were all the rage years ago (look at the commercial list at the bottom of the article).
Akamai claims that 20% of the internet's traffic flows through their network, so I'd say they're all the rage now
Time to make The Internet 2.0 (Score:2, Interesting)
#1 Every Email program requires the user to log in with a user ID and password before sending email. SMTP is too insecure and needs to be rewritten.
#2 The ISP will create an OpenID or some other universal account for the user, with the option of using up to 25 nicknames or handles instead of the real name for privacy concerns. Each OpenID will be assigned a number to keep track of the use.
#3 IPV6 will be used instead of IPV4, but there will be a portal for IPV4 users to use IPV6 IPs as a proxy via their ISP to get on the Internet 2.0 if they need to get on it.
#4 All domain names will still be owned but the web server and email hosting will only be on the Internet 2.0 if they meet with security guidelines to verify who is sending email via user ID and password. This will be to fight spam.
#5 Internet Use will be changed so that terrorist web sites, and web sites about building bombs or suicide or illegal activities will not be on the Internet 2.0 but instead be on another network called Dangernet that has to pay more money to access and has no privacy each user has to be identified for security reasons.
#6 CMS and Forum and Blog software has to have moderators or a karma system or some way to deal with trolls and flamewars. There needs to be a universal code of conduct for the Internet 2.0 but Dangernet will not have such rules.
#7 Internet 2.0 will discourage the use of pop-up ads and will fine companies for having malware infected advertising that infects user's systems. They will be fined $100 for each malware infection with the funds going to pay for repair of the Internet.
#8 Proxies will require a user ID and password to use, may not always be free, but that way they can track who is using a proxy in case of abuse. Dangernet has no such rules.
#9 The Untied Nations or some other International group will create an organization that polices the Internet 2.0 to make sure that Human Rights and Civil Rights are not being violated and there is no harassment. Warnings will be given out, and each ISP will have a TOS and if the user gets too many warnings their account will be disabled until they agree to the TOS and agree to stop the harassment or hacking or spamming or malware writing.
#10 The Internet 2.0 will create new standards for file exchange using BitTorrent but for commercial software will require TrialPay or accepting Internet offers to pay for downloading commercial software or music or videos for free.
#11 Each ISP will have a virtual machine that runs either Windows, Linux, or Mac OSX that allows their users to log in via a user name and password and use a virtual machine to get on the Internet to help protect their machine from malware.
Re:X Windows?? (Score:3, Interesting)
Heh, yeah I noticed that.
Reminds me of that cyberpunk parody http://www.netfunny.com/rhf/jokes/91q1/ozpunk.html [netfunny.com] :
I needed to read the X Windows/Motif 1.1 manual, so I came to the bar and asked Ratz to fix the documentation data in liquid form for me. It made a bitter, painful drink, but it was better than spending days turning pages in realspace.
Ratz put a bucket of liquid in front of me.
"I wanted a glass of docs, Ratz. What the hell is this?" I barked.
"Motif don't fit in a glass anymore," he barked back.
I like your sig "Anyone who loves or hates any language, platform, or manufacturer, doesn't know what they're talking about." -- there's so much frickin' negativity and mindless "can do no wrong" praise out there.
Arrogance (Score:2, Interesting)
This idiot is ... well an idiot.
The internet wasn't built in its current form!
It evolved into what we see today. The equipment used to build ARPA-net and such networks that became the internet was never supposed to be the high speed, huge subscriber numbers, massive number of application, porn-delivery system that we have today. It was made to exchange simple data. That was it.
And you can't just UPGRADE the internet (pssst... its not like in southpark and its just a big Linksys WRT54G router... really... honest!) Its a lot of moving parts.
I hate it when people with ZERO knowledge and less history try to say something is broken because their little app doesn't work correctly. Here's a hint... write cleaner, more efficient networking code!!! You'd be amazed how well that works (too bad this moron seems incapable of that).
Stupid should be painful!
I just gotta say some thing here. Damb newbies! (Score:3, Interesting)
> API Lead at Twitter, Alex Payne
Yes Newbie. He clearly has no appreciation of the history or how things came to be.
It's not like TCP/IP was the only choice, many other network technologies tried to become the Internet and failed. TCP/IP won out because it was the first to work and was open enough to work across all platforms. Novell, IBM, Microsoft all had their Networking technologies that they tried to push out TCP/IP with and almost completely did for a while with LAN's. But that all fell apart when companies want to move to WAN's and scaling up, TCP/IP was the only thing what work for both LAN and WAN. People forget the large push ATM to the desktop had, where they tried to replace TCP/IP with ATM. ATM's strength coming from preexisting telecoms switches for large voice WAN's was the only thing that supported the high bandwidth fiber for a long time. But TCP/IP just tunneled right over ATM, where ATM was too sensitive to tunnel, and was limited in what medium it can operate over. (No ATM over 2400 Baud Modems for example!)
In the end it's about Evolution, it's not some engineer or any group of humans that get to make the final decision. Call it the market, but people will choose what ever get's there job done best for them. This includes many factors that engineers never consider, legacy gear, awareness of terminology, software support, reliability, cost, platform support, open-ness of standard, multi-vendor support, cost of HW/SW. Maturity of technology, what is the TOP and and BOTTOM end.
By this I mean TCP/IP can run on a PIC Microchip and a billion Dollar super computer. It can run over radio, fiber, satellite, carrier pigeon(RFC 1149) even.
What ever takes it's place will have to be that flexible, where every light bulb can have it's own network address and still support TerraBit networks.
This is no small task, the reality is IP is flexible, so much so that you can run other protocols through it or it though other protocols. As such it will most likely be around forever and just have stuff layered over and under it. Like VPN's, PPPoE, RTP/RTSP.
Anyone is free to start creating there own IPv6 or what ever other kind of network and selling it, or running it in parallel with the internet or even over the IPv4/6 Internet.
So at this point to think your going to convince everyone to drop IPv4/6 and try something immature that is untried and untested it just unrealistic and ignorant to suggest.
Re:rough consensus and running code (Score:1, Interesting)
"We reject: kings, presidents and voting. We believe in: rough consensus and running code." - David D. Clark [wikipedia.org], former chair of the IAB
Amen! I don't think there is any other medium for the "put up or shut up" mentality than software. If you say you can design and build a better fighter plane, bridge, automobile or building you can always claim that you just lack the $$ million s to do it but you can start writing software on a old used $50 laptop and FREE tools so get coding or shut up!
Worse is better (Score:3, Interesting)
Okay, I"ve never heard of "Worse is better". The author uses it disparagingly, saying "an inferiorly designed system or piece of software may be more successful than its better-designed competitor".
But the Wikipedia article says it means Simplicity > Correctness > Consistency > Completeness, as opposed to an alternate valuing of Correctness > Consistency > Completeness > Simplicity. In other words, doing a few things right and easy is better than doing everything consistently.
I challenge his belief that doing less is inferior to doing more. Stepping away from computers for a moment, I can't think of any device that would be improved by piling on features and would not be improved by doing its key task more efficiently. I guess he's saying that the Internet isn't complex enough and can't do enough different things.