You don't seem to understand the RFC you sent me, both of which are for QoS models. "best-effort" is the default behavior of the network in the absence of traffic classification. This is the standard interpretation. If a particular QoS discipline chooses to redefine "best-effort" to mean "the lowest class", then sure, you can have "better than best effort". But that makes about as much sense as me swapping the meanings of true and false in my favorite programming language. Just because there's an RFC saying so doesn't make it true or good. Particularly since your QoS discipline applies to YOUR network. I can choose to ignore flags when your packet reaches MY network.
If you want to be pedantic, you could argue that every router has to have some QoS mechanism, because something has to decide which frame to send down the wire, and thus the QoS definition applies. But this is obviously silly, since then you could argue that a router that has no discernable traffic-shaping capability is doing QoS which then makes the "quality" part of QoS a meaningless word.
And, since you ignored the RFC (5290) that comes up first when you Google for "best-effort rfc", I'll paste the relevant portion here (which is the first sentence):
This document presents some observations on "simple best-effort
traffic", defined loosely for the purposes of this document as
Internet traffic that is not covered by Quality of Service (QOS)
mechanisms, congestion-based pricing, cost-based fairness, admissions
control, or the like.
and
Our
intention is to define "simple best-effort traffic" to include the
dominant traffic class in the current Internet.
Given that the RFC I reference is in the context of network neutrality, I think my citation is more relevant.
You might argue that "best-effort" constitutes a class of router behaviors, and that a QoS discipline makes traffic "better", thus QoS policies are "better than best-effort". But I would point out that, regardless of your QoS mechanism, that packet (or even aggregate of packets, as you point out) are subject to the same forces that make the Internet unreliable. The Internet is, at best, best-effort.
It is also very important to note that the net effect of carving networks up into bandwidth classes means that the network will be less able to handle congestion events. Why? Because network traffic is bursty. The greater your pooled resources, i.e., the greater your excess capacity, the better you are able to handle large transfers. QoS policies should be the exception, not the norm, and they are best used on private networks where the traffic distribution is understood (like your corporate LAN).
There's a great paper on pooled resources from the 1970's that I can't locate, but that gives a rather convincing argument using the probability of failure (memory allocation) given n users.
I should also add that priority queues differentiate among traffic below the best-effort threshold. Keep in mind that the total QoS from one endpoint to another is a function of the routing decisions of every node from one endpoint to the other. So Joe ISP can change the priority for a particular website, but they can only deprioritize traffic. You can't affect the policy of upstream traffic, and in most cases, there's a lot of network upstream of you, even when you're an ISP. Saying that you can "deliver better than best effort" is pure, unadulterated bullshit.
This is totally, 100% wrong. "best effort" refers to the fact that in packet-switched networks, there are NO GUARANTEES whether a packet will reach its destination. This is in contrast to connection-oriented networks, like the phone network, where guarantees can be made once a connection has been established. Wikipedia's entry on best effort delivery is also wrong: "best effort" is the upper bound on packet-switched networks. Every packet that enters a connectionless network has some probability of reaching its destination, but there is no guarantee, which is why TCP has a retransmission mechanism. If you are willing to accept occasional lost packets (because in some applications, recovering that lost data is not worth the cost), you use UDP.
Anyone who says that they can deliver "better than best effort" on a packet-switched network is blowing smoke.
It's also important to recognize that the "pooled" nature of packet-switched networks is by design, and that its loss characteristics were deemed to be an acceptable tradeoff (the advantages being cost and network neutrality).
This book has a good history on the Internet, by a person who worked on the early ARPAnet networks.
I sort of wonder if this is an olive branch from Wikileaks to the US government. After all, the United States has been pressuring Switzerland to allow investigators to peek inside Swiss accounts for awhile now. You may recall that the US offered amnesty to tax evaders using Swiss accounts who 'fessed-up a couple years ago. This provides some incentive to the US government to ease up a bit on Wikileaks.
When you're at work, you're using someone else's computer on someone else's network on someone else's time, and you're probably using a business email account that represents a business. A business has every right to read that email.
Now, a company might decide to be nice and let you check your own email and make phone calls (using their telephones) on their time, but they don't have to. I don't have any problem with this. Now if they want to look at my email on my home computer when I'm not at work, well, that's a different story.
You know, it's funny-- my parents moved us up to Maine around the time that show came out, and it really just connected with my family. I'm sure that Alaska and Maine are different in many important ways, but the weird-o small-town portrayal in Northern Exposure was totally spot-on where we lived. Ha ha. I have that on DVD somewhere... I should dig it up...
This is true. Prior to grad school (doctorate in computer science), I worked about 45/week on average, plus I took CS courses at night. I say this came out to somewhere between 60 and 65 hours per week for me.
I started working on my PhD this past fall, and I ended up AVERAGING about 70 hours a week. That extra 5-10 hours was extremely stressful, and I had a number of health problems at the end of the semester because of the exhaustion. I've also noticed that age makes a big difference here-- I'm 7-8 years older than most of my peers, and when they really want to go long stretches, I can't keep up. On the other hand, my code is usually better than theirs the first time around:)
There are patterns within patterns, however, and one of those meta-patterns is the recognition that not all patterns are valid. That's why we have the scientific method in the first place.
What's interesting is that, in modern parlance, "trailer park" has a pejorative quality to it, but this wasn't always the case. Steinbeck's Travels with Charley waxes poetic about the freedom mobile homes, even describing them as a necessity of modern life, given that jobs in the "modern era" (i.e., 1960's) were no longer as stable as they once were. I can see how we got to our present-day understanding of the term, but sometimes I think, wistfully, that Steinbeck's vision isn't such a bad one.
And, it should be pointed out, Americans can only work at LHC through a sponsoring member instituition such as a member university (like my alma mater, Boston University). Direct participation in work at the LHC (e.g., being employed there) is only available to citizens belonging to EU member nations. Shutting down Fermilab makes it even LESS desirable to become a physicist in the United States.
That may certainly be the case for the 2000-present, but what about before then? I remember using Microsoft Word on monochrome Macs. Interestly enough, I think Word 5.1a was probably the best piece of Mac software they're ever written.
My first Mac was an SE/30. I'm typing this on a MacBook Pro (and my original Apple Extended keyboard for said SE/30). I supported a large Xserve/Xsan/StorNext cluster for years. I even wrote a bootloader for the iPod mini.
If you don't have a love/hate relationship with Apple, you're not paying attention.
But also, Apple customers are used to breathlessly thanking The Steve for making their lives momentarily miserable. As long as they're promised to receive Teh New Shiny, and that all of their problems will be solved in the next iteration, Apple customers will do it.
Microsoft customers are different. Microsoft's main selling point for the last decade, whether it was true or not, was lower TCO. Bean-counters love Microsoft, and crusty old entrenched IT managers love Microsoft as well. And Microsoft's customers have poured millions into Microsoft's enterprise technology, all of which was tied closely to x86 until.NET. I know a guy who works for Philips-- when Microsoft deprecated an API that Philps had heavily invested in, Philips complained, and Microsoft un-deprecated it.
I think the main point is that Apple's important customers are largely home end-users. Microsoft's important customers are businesses. Businesses tend to be a bit more conservative.
BTW, I'm not sure the "emulated x86" platitude holds anymore. x86 chips themselves have emulated the x86 instruction set for quite awhile now. They're something like RISC inside now... Whether that means that ARM can emulate x86 well, I don't know, but it is clearly possible to do well.
What I think would be really interesting is if Microsoft decided to leverage.NET to help them expand their software ecosystem. There's already a.NET runtime on Linux: Mono. Hey... just so happens that Android runs Linux...
But I doubt that would happen. Microsoft hates making their technology available outside of the Windows environment, as evidenced by the poor quality of Mac ports of Microsoft stuff. I never really understood why Microsoft made them in the first place. They must have some calculation somewhere that shows that it's better to keep people at least partly hooked to their stuff...
Microsoft might be viewing this much the way Apple views iOS: it doesn't matter. Mobile devices, especially touchscreen devices, are different enough from their hardwired brethren that people may not seek to run the original software. Add to the fact that most technology companies are seeking to push software "to the cloud" (and indeed, Microsoft already has a cloud version of Office), this may become less and less of an issue.
I personally think that Microsoft needs to make the break to stay competitive. Continuing to support legacy software is extremely painful for both Microsoft and for their customers. I used to work for a company that was heavily invested in legacy Microsoft technologies. You know those dastardly tactics that Microsoft uses to lock you in to their product? Well, it keeps you from using new Microsoft technology as well. Loss-aversion may be irrational, but, well, you try arguing that you need to switch to new tech to a CTO who has sunk millions into software that requires ActiveX on IE6. That, my friends, is why IE6 is still around. But I'm mildly amused at the irony that Microsoft's own proficiency in the lock-in game is hurting them now.
If you're thinking that digital photography lacks the "continuous" qualities that film has, think again. All of our recording media is, at some level, discrete. The difference is that film is not digital.
Right now, large format film still has the edge qhen it comes to resolution, but I have no doubt that digital technologies will exceed even what film can offer. It will do this, mind you, while maintaining all the other things digital information has going for it (cheap to store, easy to copy, etc).
Mac GX? I don't believe there was ever such a thing. Do you mean Mac IIfx? Or maybe the Mac IIcx? Both of those were around when the Amiga was popular. The IIfx was a killer machine at the time.
The key thing is to get away from the computer for some face-to-face interaction with women (or men... I'm assuming you're a straight man here, but apologies if that is not the case). Outdoor activities like hiking groups, are strangely good places to find a partner (four of the couples I know met in hiking groups, including my sister-in-law and her husband). But if you're not an outdoorsy type, or if you think that's too dramatic a leap for you, try something like a professional interest group or a user group. Sure, these places tend to be dominated by men, but the point is to get out and expand your social network. At the very least, you're still doing something you enjoy, while having increased your chances, however slight, of meeting another person you might share an interest in.
Oh, one other thing I forgot to mention: Access has an excellent query builder. I've occasionally had to write very long queries (multiple joins, subqueries, IN or HAVING clauses, etc), and found myself getting lost while doing it. In Access, you can use the Query Designer to build them graphically, then switch to "SQL View" which spits out the SQL. It generates ANSI SQL, so you can usually just cut and paste it into code or a command interpreter. You may have to fiddle with the string quoting in psql if you're using Postgres, but mysql handles them as-is. Oh, and, as if I didn't have enough nice things to say about Access-- the Relationships window does a very nice job drawing relationships that MySQLWorkshop only recently started doing correctly (MySQLWorkshop *still* cannot auto-layout the tables without making a mess). Of course, you have to tolerate Microsoft's own variant of the ER diagram.
Slashdot Mirror
If you want to be pedantic, you could argue that every router has to have some QoS mechanism, because something has to decide which frame to send down the wire, and thus the QoS definition applies. But this is obviously silly, since then you could argue that a router that has no discernable traffic-shaping capability is doing QoS which then makes the "quality" part of QoS a meaningless word.
And, since you ignored the RFC (5290) that comes up first when you Google for "best-effort rfc", I'll paste the relevant portion here (which is the first sentence):
This document presents some observations on "simple best-effort traffic", defined loosely for the purposes of this document as Internet traffic that is not covered by Quality of Service (QOS) mechanisms, congestion-based pricing, cost-based fairness, admissions control, or the like.
and
Our intention is to define "simple best-effort traffic" to include the dominant traffic class in the current Internet.
Given that the RFC I reference is in the context of network neutrality, I think my citation is more relevant.
You might argue that "best-effort" constitutes a class of router behaviors, and that a QoS discipline makes traffic "better", thus QoS policies are "better than best-effort". But I would point out that, regardless of your QoS mechanism, that packet (or even aggregate of packets, as you point out) are subject to the same forces that make the Internet unreliable. The Internet is, at best, best-effort.
Sorry, I meant to refer to "connectionless" networks, not packet-switched. You obviously can have packets on both types of network.
It is also very important to note that the net effect of carving networks up into bandwidth classes means that the network will be less able to handle congestion events. Why? Because network traffic is bursty. The greater your pooled resources, i.e., the greater your excess capacity, the better you are able to handle large transfers. QoS policies should be the exception, not the norm, and they are best used on private networks where the traffic distribution is understood (like your corporate LAN).
There's a great paper on pooled resources from the 1970's that I can't locate, but that gives a rather convincing argument using the probability of failure (memory allocation) given n users.
I should also add that priority queues differentiate among traffic below the best-effort threshold. Keep in mind that the total QoS from one endpoint to another is a function of the routing decisions of every node from one endpoint to the other. So Joe ISP can change the priority for a particular website, but they can only deprioritize traffic. You can't affect the policy of upstream traffic, and in most cases, there's a lot of network upstream of you, even when you're an ISP. Saying that you can "deliver better than best effort" is pure, unadulterated bullshit.
This is totally, 100% wrong. "best effort" refers to the fact that in packet-switched networks, there are NO GUARANTEES whether a packet will reach its destination. This is in contrast to connection-oriented networks, like the phone network, where guarantees can be made once a connection has been established. Wikipedia's entry on best effort delivery is also wrong: "best effort" is the upper bound on packet-switched networks. Every packet that enters a connectionless network has some probability of reaching its destination, but there is no guarantee, which is why TCP has a retransmission mechanism. If you are willing to accept occasional lost packets (because in some applications, recovering that lost data is not worth the cost), you use UDP.
Anyone who says that they can deliver "better than best effort" on a packet-switched network is blowing smoke.
It's also important to recognize that the "pooled" nature of packet-switched networks is by design, and that its loss characteristics were deemed to be an acceptable tradeoff (the advantages being cost and network neutrality).
This book has a good history on the Internet, by a person who worked on the early ARPAnet networks.
I sort of wonder if this is an olive branch from Wikileaks to the US government. After all, the United States has been pressuring Switzerland to allow investigators to peek inside Swiss accounts for awhile now. You may recall that the US offered amnesty to tax evaders using Swiss accounts who 'fessed-up a couple years ago. This provides some incentive to the US government to ease up a bit on Wikileaks.
When you're at work, you're using someone else's computer on someone else's network on someone else's time, and you're probably using a business email account that represents a business. A business has every right to read that email.
Now, a company might decide to be nice and let you check your own email and make phone calls (using their telephones) on their time, but they don't have to. I don't have any problem with this. Now if they want to look at my email on my home computer when I'm not at work, well, that's a different story.
You know, it's funny-- my parents moved us up to Maine around the time that show came out, and it really just connected with my family. I'm sure that Alaska and Maine are different in many important ways, but the weird-o small-town portrayal in Northern Exposure was totally spot-on where we lived. Ha ha. I have that on DVD somewhere... I should dig it up...
This is true. Prior to grad school (doctorate in computer science), I worked about 45/week on average, plus I took CS courses at night. I say this came out to somewhere between 60 and 65 hours per week for me.
:)
I started working on my PhD this past fall, and I ended up AVERAGING about 70 hours a week. That extra 5-10 hours was extremely stressful, and I had a number of health problems at the end of the semester because of the exhaustion. I've also noticed that age makes a big difference here-- I'm 7-8 years older than most of my peers, and when they really want to go long stretches, I can't keep up. On the other hand, my code is usually better than theirs the first time around
There are patterns within patterns, however, and one of those meta-patterns is the recognition that not all patterns are valid. That's why we have the scientific method in the first place.
Oh, I just thought it was correcting for the inevitable mispronunciation that we're going to do in our heads.
What's interesting is that, in modern parlance, "trailer park" has a pejorative quality to it, but this wasn't always the case. Steinbeck's Travels with Charley waxes poetic about the freedom mobile homes, even describing them as a necessity of modern life, given that jobs in the "modern era" (i.e., 1960's) were no longer as stable as they once were. I can see how we got to our present-day understanding of the term, but sometimes I think, wistfully, that Steinbeck's vision isn't such a bad one.
And, it should be pointed out, Americans can only work at LHC through a sponsoring member instituition such as a member university (like my alma mater, Boston University). Direct participation in work at the LHC (e.g., being employed there) is only available to citizens belonging to EU member nations. Shutting down Fermilab makes it even LESS desirable to become a physicist in the United States.
That may certainly be the case for the 2000-present, but what about before then? I remember using Microsoft Word on monochrome Macs. Interestly enough, I think Word 5.1a was probably the best piece of Mac software they're ever written.
My first Mac was an SE/30. I'm typing this on a MacBook Pro (and my original Apple Extended keyboard for said SE/30). I supported a large Xserve/Xsan/StorNext cluster for years. I even wrote a bootloader for the iPod mini.
If you don't have a love/hate relationship with Apple, you're not paying attention.
But also, Apple customers are used to breathlessly thanking The Steve for making their lives momentarily miserable. As long as they're promised to receive Teh New Shiny, and that all of their problems will be solved in the next iteration, Apple customers will do it.
.NET. I know a guy who works for Philips-- when Microsoft deprecated an API that Philps had heavily invested in, Philips complained, and Microsoft un-deprecated it.
Microsoft customers are different. Microsoft's main selling point for the last decade, whether it was true or not, was lower TCO. Bean-counters love Microsoft, and crusty old entrenched IT managers love Microsoft as well. And Microsoft's customers have poured millions into Microsoft's enterprise technology, all of which was tied closely to x86 until
I think the main point is that Apple's important customers are largely home end-users. Microsoft's important customers are businesses. Businesses tend to be a bit more conservative.
BTW, I'm not sure the "emulated x86" platitude holds anymore. x86 chips themselves have emulated the x86 instruction set for quite awhile now. They're something like RISC inside now... Whether that means that ARM can emulate x86 well, I don't know, but it is clearly possible to do well.
What I think would be really interesting is if Microsoft decided to leverage .NET to help them expand their software ecosystem. There's already a .NET runtime on Linux: Mono. Hey... just so happens that Android runs Linux...
But I doubt that would happen. Microsoft hates making their technology available outside of the Windows environment, as evidenced by the poor quality of Mac ports of Microsoft stuff. I never really understood why Microsoft made them in the first place. They must have some calculation somewhere that shows that it's better to keep people at least partly hooked to their stuff...
Microsoft might be viewing this much the way Apple views iOS: it doesn't matter. Mobile devices, especially touchscreen devices, are different enough from their hardwired brethren that people may not seek to run the original software. Add to the fact that most technology companies are seeking to push software "to the cloud" (and indeed, Microsoft already has a cloud version of Office), this may become less and less of an issue.
I personally think that Microsoft needs to make the break to stay competitive. Continuing to support legacy software is extremely painful for both Microsoft and for their customers. I used to work for a company that was heavily invested in legacy Microsoft technologies. You know those dastardly tactics that Microsoft uses to lock you in to their product? Well, it keeps you from using new Microsoft technology as well. Loss-aversion may be irrational, but, well, you try arguing that you need to switch to new tech to a CTO who has sunk millions into software that requires ActiveX on IE6. That, my friends, is why IE6 is still around. But I'm mildly amused at the irony that Microsoft's own proficiency in the lock-in game is hurting them now.
No, it is not digital. It is discrete. Being discrete is certainly one of the properties required to be digital, but it lacks some others.
If you're thinking that digital photography lacks the "continuous" qualities that film has, think again. All of our recording media is, at some level, discrete. The difference is that film is not digital.
Right now, large format film still has the edge qhen it comes to resolution, but I have no doubt that digital technologies will exceed even what film can offer. It will do this, mind you, while maintaining all the other things digital information has going for it (cheap to store, easy to copy, etc).
Hey, man, soft porn is the positive* view of the future.
* for certain values of "positive"
Mac GX? I don't believe there was ever such a thing. Do you mean Mac IIfx? Or maybe the Mac IIcx? Both of those were around when the Amiga was popular. The IIfx was a killer machine at the time.
Maybe so, maybe so. All is know is that 1720 was a really bad year.
The key thing is to get away from the computer for some face-to-face interaction with women (or men... I'm assuming you're a straight man here, but apologies if that is not the case). Outdoor activities like hiking groups, are strangely good places to find a partner (four of the couples I know met in hiking groups, including my sister-in-law and her husband). But if you're not an outdoorsy type, or if you think that's too dramatic a leap for you, try something like a professional interest group or a user group. Sure, these places tend to be dominated by men, but the point is to get out and expand your social network. At the very least, you're still doing something you enjoy, while having increased your chances, however slight, of meeting another person you might share an interest in.
Oh, one other thing I forgot to mention: Access has an excellent query builder. I've occasionally had to write very long queries (multiple joins, subqueries, IN or HAVING clauses, etc), and found myself getting lost while doing it. In Access, you can use the Query Designer to build them graphically, then switch to "SQL View" which spits out the SQL. It generates ANSI SQL, so you can usually just cut and paste it into code or a command interpreter. You may have to fiddle with the string quoting in psql if you're using Postgres, but mysql handles them as-is. Oh, and, as if I didn't have enough nice things to say about Access-- the Relationships window does a very nice job drawing relationships that MySQLWorkshop only recently started doing correctly (MySQLWorkshop *still* cannot auto-layout the tables without making a mess). Of course, you have to tolerate Microsoft's own variant of the ER diagram.