|
The Air Force employs 275,000 civilians and contractors.
Their VPN network can only support 72,000 simultaneous connections.
That explains a lot...
".45 ACP - because shooting twice is just silly" - JSOP, 2010
-----
You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
-----
When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
I would have thought, pfft, 72,000'l be enough. 
|
|
|
|
|
It's not the pipe, it's what they're smoking.
|
|
|
|
|
640K (connections) will be enough for everyone!
Software rusts. Simon Stephenson, ca 1994. So does this signature. me, 2012
|
|
|
|
|
56K connections should be fast enough for everyone!
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
OriginalGriff wrote: 56K connections should be fast enough for everyone!
God, yes!
I was totally blown away, when I upgraded from 28k!
... But then along came javascript...
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
It is remarkable, given how far CPU power, memory size, and bandwidth have come since the days of the Apple ][+, that mindless designs still cause some things to run like pigs in $#!+.
|
|
|
|
|
It is also remarkable how clever we are at abusing resources "just because they are there", our senseless expectations and our demand for resources that we do not need at all.
An example of the latter: I've got 16 GB of RAM my home computer. Occasionally, I have seen 8 GB in use. Yet several of my friends claim that they need 32 GB to get the maximum speed out of the machine. I ask to have a look at the RAM load as we are sitting there - it may be in the 4-6 GB range. Yet they insist: If I simultaneously run video editing and a sound editor and my wife's recepie database and ... How often do you? Well, I am saying if I do! ... It has happened that by starting all the big applications they've got on their PC, that they have managed to cross 16 GB, barely. "There you see! I need 32 GB!".
One of my coworkers, a computer guy, was complaining that his computer was sluggish, lots of paging. When he counted his Chrome tabs, there were 120 of them, some of which he hadn't opened for a month. If he restarted Chrome, it would take him a lot of work to set up all the tabs again! But to plug in more RAM he would have to do that anyway. So he restarted Chrome and haven't complained about the speed of his machine since.
Senseless expectations: I was responsible for a build configuration management system, that would set up the right tool and package versions for a build. You could run it interactively on your desktop PC to verify that it was correctly set up. One of my coworkers rejected it completely, it was unbearingly slow. I timed it, with his tool profile (well above a hundre different components); the check took less than two seconds. "But that means I have to wait for it to complete! It should reply immediately, in order to be usable!"
Along the same line: My internet connection is 100 Mbps; I can download a full 4.7 GByte DVD in less than ten minutes. Yet lots of people think a 1 Gbps connection is worth its money, so they can have the download completed in less than one minute. After spending 20 minutes searching the catalog, and I will spend an hour and a half to watch that movie later, I dont give a dime whether the download takes one or ten minutes! That is a certainly not a time-critical, interactive operation! (Besides, to utilize my 100 Mbps connection at a sustained 100 Mbps, I must start at least two large transfers in parallel; few if any server will deliver that speed alone, and I doubt that any server will deliver 1 Gbps to you across a WAN Internet.) Yet you see lots of people complaining about the authorities considering any optical fiber a "broadband" connection: They insist that 100 Mbps is the very minimum speed qualifying as broadband!
|
|
|
|
|
I was going to make a silly statement: surely not all of those 275,000 need to be on the network. Then I realized how silly it was. Practically all personnel, regardless of the enterprise, need network access these days.
Example: There's a housing development going up on (what was) my usual route to work. I've noticed a lot of the workers carrying around tablets or phones. The temporary utility stand on the corner has some orange cables coming to it I suspect are fiber optic carrying broadband Internet, and there's a box with a couple antennas that's probably WiFi.
We haven't run out of VPN bandwidth yet where I work, but I've copied a crap-ton of stuff to my home laptop this weekend in case it goes belly-up.
Software Zen: delete this;
|
|
|
|
|
We are handling different kinds of resources: The number of simultaneous VPN connections applies even if each connection transfers a hundred bytes a minute or less - it is a function of the table size, or whatever kind of data structure the VPN software uses to manage the connections. A larger table does not in itself imply a higher bandwidth load.
It is not given that every user needs to be connected through a VPN. E.g. my company provides a lot of documentation through ordinary web pages. Lots of workers may have to look up documentation, or possibly make company database lookups for non-sensitive data, where a simple password/TFA protected login provides enough protection.
I do most of my work through a TFA remote login. Our old VPN couldn't support enough connections to send everybody home, but we had a replacement in the trial phase that was planned to last for another few weeks. They decided to accelerate the move into ordinary operation, and it seems to handle it well, and next week I will start using it. But even if it had failed, I still could do a lot of work through non-VPN channels.
For the bandwidth question: I must say that I am quite impressed by how well the backbone networks handle this! Of course response is sluggish in rush hours, but we have had no major breakdowns, and video connections are still not as jerky as in the ADSL days  . Most likely, the providers will learn a lot from this, so they can optimize routing strategies and capacities for the next similar crisis. (E.g. if you do a traceroute, you may be suprised how many hundred miles the packet travels to cross the road. A few years ago, here in Norway every packet that did not have origin and destination with the same ISP went via an Internet Exchange in Oslo, which might be a 3000 km round trip. Today, there are exchanges in a handful towns, but I guess the current experience will give valuable input to new locations where such exchanges can reduce backbone load.)
|
|
|
|
|
In order to telework, they all have to be able to connect, and as far as I know, it has to be over VPN. However, a certain number of people are still going in to work (essential personnel). However, that probably leaves at least half of the workforce trying to use VPN, AND THIS DOESN’t even consider the military personnel that need to use VPN.
It’s a mess...
".45 ACP - because shooting twice is just silly" - JSOP, 2010
-----
You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
-----
When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
Every now and then I get the urge to publish a book on the pearls we left behind.
Solutions to problems that we are fighting with today - but ten, twenty or fifty years ago, we did have solutions.
At the time, they may have been rejected because computing resources where not available. Today, they are, yet we haven't picked up the old solutions.
We may have settled on architectures where the old solutions do not fit in. So we can see what would be the solution, but we cannot make use of it.
Some solutions may not have been viable in its day - that's why it was left behind - and maybe not even today. Yet it might represent interesting concepts or approaches, that could be mind stimulating today, even though not necessarily realizable as code.
Often, I see that my own thoughts have spun further: This is a great basic concept, but it could have been expanded so-and-so. That certainly is no discredit to the concept - quite to the contrary, if your ideas to extend it could have made it even stronger.
Some concepts "sort of" exist, mostly in papers and articles, but are not commonly practiced in coding.
I've got a fair number of candidates of my own, but I'd like to hear which other concepts, techniques and models you think that today does not receives the merit that it deserves. What is your favorite forgotten concept, that we really should be aware of today?
If I really sit down to write this book, I probably will not see offhand what makes it valuable anno 2020, so I might need your explanations and references. We could start it here, and move it elsewhere if it grows too large.
Would you think it interesting to start this a "book" as a CP article, that we could develop stepwise together?
|
|
|
|
|
The concept that all data would include semantic meaning and the things you could do with it would be determined by plugins. This was the original vision of the GUI created by the folks at Xerox PARC, and is something I consider forgotten but that we should really be aware of today. Instead, we have monolithic OS's where the semantic nature of information is totally lost in bit streams and meaningless JSON packets. 
And I've already written a variety of articles here on CP to explore the concepts of true semantically aware data. 
|
|
|
|
|
Many of the concepts in my book and articles on this site are more or less in that category. Maybe they haven't been left behind so much as they aren't widely known. If you drew up a list of topics, I'd pitch in if I thought I could contribute something useful.
|
|
|
|
|
Greg Utas wrote: Many of the concepts in my book and articles on this site are more or less in that category.
Grandiosity points awarded.
«One day it will have to be officially admitted that what we have christened reality is an even greater illusion than the world of dreams.» Salvador Dali
|
|
|
|
|
Every month that has passed since a new concept was introduced halves the number of point awarded to any reference to that concept. So a year-old concept is worth 1/2^12, or quarter of a per mille, of a concept developed this month.
|
|
|
|
|
Digital concrete?
Sounds like something I'd add to my kilometre-high "to read" pile.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Member 7989122 wrote: Would you think it interesting to start this a "book" as a CP article, that we could develop stepwise together?
Would I find interesting any further discussion of this rambling, content-free,
Member 7989122 wrote: on the pearls we left behind.
fantasia ?
No.
«One day it will have to be officially admitted that what we have christened reality is an even greater illusion than the world of dreams.» Salvador Dali
|
|
|
|
|
OK, so I understand that you think there is nothing to learn from the past. Anything that has ever been rejected, for whatever reason, under whatever assumptions and circumstances, should forever be forgotten and ignored, and noone should ever suggest that it might have some merit under the new and changed circumstances. A rejection of an idea should always be final and absolute.
Fine. I'd like you think so - that is your right. So I am certainly NOT going to bother you with any stuff about ideas that you have been decleared, for yourself, as rejected and carrying no value. Obviously, I cannot prevent you from looking up, or by accident come across, any such idea that you consider rejected and without value, but I certainly feel no need to force it onto you. You go ahead hunting the freely roaming game, whether it be the web framework of the week, the next programming language to hit the Tiobe index as the fastest growing one, or whatever. Anything else is just soooo 2019, and not worth a dime!
modified 23-Mar-20 21:57pm.
|
|
|
|
|
"Never wrestle with a pig. You both get dirty and the pig enjoys it." --apocryphally attributed to George Bernard Shaw
|
|
|
|
|
Member 7989122 wrote: OK, so I understand that you think there is nothing to learn from the past. Anything that has ever been rejected, for whatever reason, under whatever assumptions and circumstances, should forever be forgotten and ignored, and noone should ever suggest that it might have some merit under the new and changed circumstances. A rejection of an idea should always be final and absolute.
wonderfully entertaining hyperbole, but, you left out the part where I strangle kittens 
«One day it will have to be officially admitted that what we have christened reality is an even greater illusion than the world of dreams.» Salvador Dali
|
|
|
|
|
... "Aisle B, Back".
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
There will be a free public seminar on the topic of "Avoiding crowds during the coronavirus pandemic." Everyone is encouraged to attend. 
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
|
Yes, I thought it was a joke, too[^].
And this[^].
And at least a dozen others.
It's reminiscent of what happens if you tell your granny that the format command is used for cleaning discs.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
