83! It's easy to remember, simple, and comes default in 100% of linux distributions.
Of course, there are those who are type 82 bigots. I can see how that's important, but with RAM prices so low these days....
Matt
Re:How do you think companies will react?
on
New ICANN TLDs Are Live
·
· Score: 3, Informative
I work for the company managing the technical backend for.pro, and while I understand you're saying this at least partially tongue-in-cheek, I need to correct your misconception..Pro is for licensed professionals onle - at the time of veritifcation, license numbers, etc. will be submitted. There will be subdomains of.law.pro,.med.pro, etc.
Now, whether that's useful to anyone, I don't know. But no finnair.pro, I'm afraid.
Thanks,
Matt
Note: I work for a domain registrar, but I don't speak for one.
.ORG was never intended to be for just non-profit organizations, and has never been enforced as such, except perhaps at the registrar level by ignorant registrars. According to RFC 1591 by Postel himself:
ORG - This domain is intended as the miscellaneous TLD for organizations that didn't fit anywhere else. Some non-government organizations may fit here.
See - its miscellaneous. In addition,.ws is a country-code TLD - in this case its for Western Samoa. The concept of.ws being for a "website" is purely marketing.
As far as polluting the namespace - the DNS system is designed to support a large number of TLDs. The restraints that need to be placed are policy ones, not technical ones.
Thanks,
Matt
Note: Although I work for a domain registrar, I don't speak for one.
You obviously haven't read my post closely - ICANN doesn't want to limit the number of TLDs, but it would be foolhardy to open the floodgates without at least giving a testbed to examine the implications of adding TLDs. As for them taking years to do this, it would be just as foolhardy to fundamentally change the structure of the internet without examination and testing of changes they'd like to make - incremental movement rather than revolution.
Have you dumped the root zone lately? The root zone is only about 60k in size. Each TLD only requires about six records - a couple of NS records for the TLD, plus the glue A records. the GTLDS(.com,.net, and.org) are on the gtld servers, and they're the ones with the 2.5 gig zone file. The load on the root servers will increase a negligible amount if 200 new TLDs were added tomorrow.
People really don't understand how the process that the Internet is going through with regards to new TLD additions. This is a TESTBED phase - the whole concept being that they try adding several new tlds with varied scope and purposes to see what the implications and results are. This is the same process that was done when Network Solutions was the only domain name registrar. 5 testbed registrars were chosen to see what would happen when the market was opened to competition - one year later, the market was opened up as a whole and today there are several hundred accredited registrars, and one of the top 3 registrars (Tucows/OpenSRS) was not even one of the testbed registrars.
The reason the testbed phase is important is because of exactly the reasons people are complaining about the new TLDs. Now that people have noticed that there's a problem with trademark holder verification, perhaps when the process is opened up (which ICANN has said it will, assuming the testbed phase works) that will be remedied. You have to look at this as a feasibility experiment. Look at the new TLDs:
-.pro - restricted use, but unique in that it provides identity/professional proof of id (for lawyers, etc)
-.info - unrestricted tld, just like.com. Come and get 'em
-.biz - for businesses only, iirc. Semi-restricted TLD
-.museum - very specific restricted TLD
-.coop - for non-profits, etc.
This is a textbook example of what should do for a feasibility study - select examples of each type and put them into production. See what happens. Make note of what works and what doesn't and use that to formulate an overall policy.
The idea that ICANN is somehow for limiting the number of TLDs is ludicrous - everyone, from the internet populace at large up to domain name registrars, want new top-level domain names. Everyone would win. But ICANN cannot simply open up the field without understanding and learning about what the implications are. People are looking almost entirely at the technical issues with adding new TLDs, while completely ignoring policy and procedural issues. Issues like false trademark submissions only prove how necessary this process is.
The other issue with Fibre Channel is that its simply SCSI-III encapsulated in Fibre Channel Packets, so there's a lot of overhead.....which makes Fibre not quite the magic bullet that people say. A fibre channel link is only good for maybe 70 MB/sec. because of signaling and overhead. But, on the plus side, You can take your fibre channel loop, stick your servers at the ends, and run IP over Fibre Channel - you just need to make sure your cards support is. I don't think any do under linux, but under Sun most do.
Thanks, Matthew J Zito, CCNA
Re:The Cost of Maintaining an Archive
on
Deja For Sale
·
· Score: 1
Yes, yes, yes.....the cost is that low when you look at the simple cost of the drives. Then you need to add up all the other costs:
-Salaries of people to run it: 1 million/year (20 people at an average of 50,000/year each, which is low)
-Database software: Assuming you want to run oracle, that'll cost you a variable amount depending on the number of cpus you want to run it on - say a 16-cpu box...... 400,000
-Storage system - I know the cost of drives are 50,000, but you need the software, the support infrastructure - 10 terabytes of storage doesn't just show up in a box. You need something like an EMC. Let's say it is an EMC, and you get a good deal - That'll be about 8 million dollars (you may think I'm kidding or exaggerating, but I'm not)
-Database server- this database won't run on a linux box - you're talking something like Sun. I don't know IBM or HP boxes, so I'll price out a Sun - a 16-cpu, 16gb of ram sun e6500 will run you about 600,000 and you need two of them (at LEAST) - 1.2 million
-HA software - Stuff like Veritas, your backup software, software raid, cluster server, etc. - 400,000
Now we're up to 11,000,000 dollars, and that's initial costs, not counting co-lo space, front-end application servers, backup servers, bandwidth, office space, benefits, management, etc. etc. etc. Of course, if you're a pre-existing company, you may be able to leverage some of your existing infrastructure to support this, but no matter what, there's no way this will cost 50k.
At my work, we run two large databases, one that's about 95 gigs and the other 160 gigs on linux. Now, we run our production db on solaris, but the data warehouse and the ticketing db are on linux with oracle. We've had great results with linux for the most part - the biggest problem is that the documentation by Oracle is not as good for linux as it is with Solaris, and its harder to find DBAs with Linux experience.
I can't vouch for Sybase's stability under linux, but Oracle will do you just fine. Get a dual or quad-cpu box, depending on how much data you need to do, and 2 gigs of RAM either way.
Well, when you're at the 30 gigabyte size, your options as far as remote replication are a little limited. For example, its not enough to warrant the sort of large-scale storage array that an EMC Symmetrix would offer that comes with built-in remote replication (SRDF).
What you could use is Veritas Volume Replicator. It runs as a service/daemon on your box and mirrors every write over IP to another box. It can be configured to do it synchronously (the I/O blocks until the remote I/O completes) or asynchronously (higher performance because there's no delay, but you run the risk of data loss when the db server goes down).
Unfortunately, Veritas VR is not available under linux - I think you said it wasn't under linux anyway, but a lot of people are offering linux solutions.
Also, given that your database is only 30 gigabytes, do you actually do a lot of writes? Realistically, if you only do a couple of hundred inserts an hour, you could just, every hour, manually insert the changed records into the remote db. Heck, do it every 5 minutes. I'm not familiar with Sybase, but on Oracle, you can just run the redo logs on the remote data center. That's going to be the cheapest option, and the most linux compatible.
Anyway, if this is really enterprise-level, spring for Veritas - their stuff is expensive but really good.
My GOD! I see it all so clearly now! Roblimo HIMSELF has been assimilated! Who's next? CmdrTaco? OR..dare I say it..JON KATZ!
We all need to send shock troops right away to the Geek compound.
In all seriousness, though, this article raises issues that I think most of the Slashdot readership are aware of, like the fact that although Microsoft may have questionable business practices, their coders are real human beings, many of whom are just trying to have a normal programmer existence. Its only the trolls, flamers, and l33t ones that somehow delude themselves (or act like they do) into believing that Microsoft is roughly equivalent to the 3rd circle of Hell.
Now, something else I find interesting that isn't brought up is whether these trolls and flamers are hurting the linux community as a whole. We've seen in the past where people have posted negative linux articles on places like ZDNet and been flamed up and down. Next time they have to write an article, do you think they'll pick something about linux, knowing that they could spend a fair chunk of the rest of their day just sorting out the flames? Or will they just write an article about NT, or Sun's newest upgrade to their Starfire system ?
This doesn't seem to happen as much (at all, really) with Sun articles, or Oracle articles, or even Windows articles. I think that's because the grass-roots, "us-against-them" mentality that has spurred a lot of really great linux development, press, and most importantly users has also created a jihad in the minds of some people who see anything anti-linux as a death knell for linux. I think linux has been raised to a certain public awareness, though, where we can take some bad articles, get the useful information out of them, and move on. If its FUD, fine. Everyone spreads FUD about everyone else, which doesn't make it okay, but DOES mean that its not some microsoft conspiracy.
Which, after a long tangent, brings me back to the original post.
I work for one of the major registrars - NSOL simply runs the SRS (Shared Registry System). Your domain contract is with your registrar. All the registrars sign a contract with ICANN regarding domain disputes, etc.
Well, but the point is that the prevailing theory among genealogists was that any given surname originated from a number of people - this suggests that that may not be the case.
The circuit he is talking about is not a 56k circuit. The 56k P-t-P you are talking about is a DS0 w/ 8kb reserved for signaling (64kb - 8kb).
An alarm pair is (IIRC) a lower-bandwidth P-t-P circuit that is routed differently through the Telco network. Wish I remember more specifics about it, but I've never ordered one myself. There's no QoS guarantee for circuit noise, etc. Caveat Emptor.
Regarding the actual implementation of this, there was a discussion on this topic on both the ISP-EQUIPMENT and Cisco GroupStudy mailing lists. The general consensus was this:
1) Make sure the distances from the PoP to the CO is less than 12,000 feet for the best chance. 2) Not all Bells will allow you to get these unloaded copper pairs, and those that do will often not give you the highest quality circuits.
I've never done it myself, but there seems to be a fair amount of success in the matter in the marketplace.
Matt
Matthew J Zito, CCNA
Re:Even other BSDs are choking off BSDI.
on
Free Solaris 8
·
· Score: 1
BSDI for a while had a speed niche. The TCP/IP stack on BSDI years ago was really fast, and BSDI was the first (IIRC) BSD unix to support SMP. Plus, it came with good tech support and a pretty decent set of software when you bought it. It was a nice package.
Of course, that was in 1997.
Since then, everyone beat them into the ground. This is an almost-verbatim transcript of a conversation between me and a BSDI engineer about 6 months ago: Me: So, do you support the IBM Netfinitys? Engineer: No. M: Do you support Fibre Channel? E: Maybe by the third quarter of 2000. M: Do you support Gigabit Ethernet? E: No. M: Do you support clustering or shared arrays? E: No.
(3 weeks later, I call the guy back)
M: Hey, we've fired up BSDI on a netfinity, but the 10/100 ethernet card in it will only run at 10 megabits. Its an AMD PcNet 100. E: Oh, that driver only supports 10 megabits. M: Any timeframe on when that'll be fixed? E: I couldn't say.
Good point. My code was snagged from the response from the DVD CCA linked to on slashdot earlier in the day. Since its a legal document, I feel fairly safe in saying its genuine. Here's the md5 hash:
5541139af5ed9ac22e855c30e6927f54
Again, if anyone gets a better source, send it to me!
I find it very interesting that some sort of serious legal authority is getting involved in this situation - I was under the impression that in the U.S. it was a civil suit. Is it a civil case in Norway also? Who are these people? Can someone fill that in?
This was a wake-up call for me. I was supportive of the EFF and everyone else involved in the suit in California, but I figured it would be one of those things where the case went to court, the judge laid the smack-down (so to speak) on the plaintiff, and we all went on our merry way. But the fact that an indictment was returned against this gentleman shows that not only is the DVD CCA intent on making a serious effort to put a stop to legal and legitimate reverse engineering but that there is a severe possibility that WE MIGHT LOSE.
I went and read the response by the DVD CCA- their argument is since the code has the master key in it, it MUST have been obtained illegally. This is a circular argument. But if they are able to convince a judge that this is true, this could signal an end to the idea of black-box reverse engineering.
How can we prevent this from happening? I'm starting by putting a copy of the DeCSS code up on my personal web server: http://128.122.106.158/decss.txt This is the only code I have - if someone wants to email me something more complete, I'll put that up. Email me at matthewzito@yahoo.com.
Everyone should put this code on their site. If enough people put this code on a website somewhere, the DVD CCA can't sue/arrest/harrass everyone. It's an old, hokey protest tactic, but it works.
Next, contact anyone and everyone in power. Call newspapers, politicians, and tell anyone and everyone who will listen. If they won't listen, tell them multiple times. Be polite, but be firm.
Write letters to the editor. Here, the idea is to make sure that everyone is aware of the issue, and more importantly, is aware of our side of the issue. Make sure that if someone reads an article that supports the DVD CCA, they have already read or heard something from someone sympathetic to our cause.
Donate money to the EFF, and any legal defense fund that is created for this gentleman. It doesn't have to be a lot, but anything you can give can help insure that we get a victory in the courts.
Anyone interested in putting together a little fund to run an ad in a major US newspaper like the NY times with the DeCSS code in it? That's more of a farfetched idea than most of the others in this post, but its a beautiful idea nonetheless.
Basically, the only things we can do are practice social disobedience (even in such a minor form as keeping our own public copies of the decss code), support those who are fighting the legal battles in the courts, and educate, educate, educate. Even if we lose some court battles, if the public in general is aware of the issues at stake, that gives us an advantage for future confrontations with companies trying to stop legitimate reverse engineering. Which there will certainly be.
Well, this is true to a certain extent, but there's some more details that need to be filled in:
Corporations are individuals under the law. In his original post, Mr. Rideau uses the word "companies". There are two general types of companies: Proprietorships and Corporations. Proprietorships are the most common form of companies. The different types of proprietorships are the Sole Proprietorship, the General Partnership, and the Limited Partnership Under a proprietorship, there is no separate legal entity- the owner's possessions and assets are open to legal action. In that situation, the transfer of data from one employee to another constitutes a transfer of the so-licensed software from one individual to another. Therefore, either individual would have the right to re-distribute that software as they saw fit. Even a non-disclosure agreement would not absolve one from this responsibility, since the GPL itself says (emphasis mine):
7. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues), conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all. For example, if a patent license would not permit royalty-free redistribution of the Program by all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Program.
So, based on this, if one is employed at a company that is a sole proprietorship (which is the most common type of company in the US) and is bound by a NDA not to release or discuss the source code to the public, they are not allowed to redistribute any of the changes back into the company, since that would be distributing the source code in violation of the above conditions.
Now, while a corporation IS its own entity under the law, there is something called "Respondeat Superior", or "the acts of the agent are the acts of the owner" (that information, and the information in general regarding rights of agency courtesy of "Small Business Advisor" pg. 446-449). If an employee of a corporation starts using GPL'ed software, and does so on company time, is the company then bound under the GPL? It depends. If the person in question either has a) a clearly stated authority to speak on behalf of the corporation (i.e. in the articles of incorporation) or b) an implied authority to do so (like you're the head of the software design department and the CTO just said, "We need you to pick a new piece of software to build"), then the corporate entity is bound by the actions of the individual.
Therefore, if a someone within the company who meets neither of the above conditions takes a piece of GPL'ed software, modifies it, and then distributes it to five other employees, s/he is bound by the GPL, NOT the company. OTOH, if someone with authority starts using a piece of software, the company is then the entity bound by the GPL.
Based on that, since Mr. Stallman said that redistribution within a company is not redistribution in terms of the GPL (we have to assume that he means a corporation, since its been demonstrated that transactions within a proprietorship constitute transactions between individuals), it is legally possible to create an organization like the one Mr. Rideau described, where upon entering the corporation you agree to not distribute internal corporate information. Since it is not redistribution according to Mr. Stallman to distribute within a corporation, then section 7 of the GPL with regards to anti-distribution and restriction thereof does not apply, since a new instance of the GPL has not been created. All the employees of the corporations consitute a part of that single entity, so their individual activities within the corporation can be managed.
The upshot of all of this is that Mr. Rideau is right. Acceptance of the GPL by a person in a corporation with the right to do so binds the corporate entity to the GPL. Since internal distribution of the code is allowed, employees can be bound legally not to distribute outside the corporation, since they have not been bound under the GPL - the corporation has.
Imagine a situation where someone takes a piece of GPL code, say Linux, as the most obvious example in this forum. A corporation is formed, and the CEO accepts the GPL. The corporation makes a new, improved version of linux (for the sake of an example, we'll say they rewrite the memory handling code and double speed). Normally, if they wanted to redistribute, they would have to make public their changes. Instead, though, they make the software available by selling small stakes in the company (or simply on a fee basis and making them part of the corporation). So, if you want the software, you have to become part of the corporation. Once you are part of the corporation, you're bound by the rules of the corporation regarding redistribution, not the GPL (since the GPL does not apply to internal corporate redistribution). If the company wanted to release the code to anyone outside of the corporation, then they would have to make it public. But since they're only giving it to other people within the corporation....
Now, I realize that this is a remote situation. But its important that we figure out these weaknesses in the GPL before someone else does and exploits them.
Did you read the post? They're going to scan for proxy servers that are open. If a user has a properly configured proxy server, there should be no way to distinguish that machine from a normal machine hooked up to the cable modem.
I don't buy into the "only gurus should be messing with such mystical things" argument, but at the same time, if someone doesn't configure their software properly, its perfectly acceptable for someone to step in and say, "Because your software is indirectly contributing to a problem, we are going to put a stop to it". Hopefully then the user will step back, find the problem, reconfigure it, and everyone will be better off. The blame isn't on @home for this. Yes, they probably could have been more on the ball about dealing with the customers who were spammers, but as an @home customer, I know a) how many people are running WinGate without really having a strong grasp of the issues involving that and b) how many portscans float by my machine on a day to day basis.
I think the UDP was warranted, and I believe that the response from @home was also. It is people with unsecured proxies that contribute to the problem, not "anyone using a home network". As I said previously, the only ones who will be affected by this scan are those whose proxies are incorrectly configured.
I remember that because the freebsd main ftp site got hammered the last time Slashdot reported right away that a FreeBSD release was out, the FreeBSD core team asked slashdot to hold their posts of new releases until the ftp mirrors catch up.
One of the problems with consumer Internet technologies (i.e. ISDN, SDSL, etc.) is that there is less of a QoS guarantee. In my area, if your ISDN line goes down, Bell Atlantic tries to get a service technician out in a few days, generally more like a week. When you're a small company starting out that could kill you. By contrast, when you have a Frame Relay (also sometimes inaccurately called a Burstable T-1) or T-1 link, Bell will immediately look into a service failure. While they're still not responsive, there is a much higher commitment to service for the higher-end data services. You also get a day of service free when there's an outage that lasts longer than 20 minutes.
My.02 about the whole issue RE: Colocation vs. home data line is that colocating will always be cheaper, given two companies of similar quality. That is, colocating with Exodus (excellent NYC colocation provider) is about $1000/month per megabit. By the same token, a T-1 through a high-quality provider like Sprint will run about $1100-$1600/month (depending on the circuit charges). You can get really crappy colocation for as low as $100/month.....just like everything else, you get what you're willing to pay for. Definitely shop around. And ask questions like:
1- What hours can I physically work on my server? 2- What kind of technical support do I get? 3- Do they offer remote power management (some colo places give you a web page where you can power cycle your box- its great)? 4- Traffic analysis? 5- Will my server be connected to a hub or a switch? (cheap places tend to hook you up to a cheap hub- this will degrade your network performance and allow people at other colocated servers to sniff your traffic)
Ask these questions and in an old Ask Slashdot there was a thing about co-location providers...go dig that up. There was a lot of really useful stuff in there.
Well, I asked my eye doctor about the surgery about 6 months ago, when I had my last exam. He said I'm an excellent candidate, etc. etc. etc. but then talked about his misgivings about the procedure:
1) He doesn't recommend it until you're 25- he says that before then your eyes are still growing and changing.
2) There have been no long term studies on the after-effects of the surgery. He described how before the surgery, the cornea is a grid of perfect hexagons, and after the surgery, the grid is gone, replaced by fused cells, small lesions, etc. He said that theoretically that should have no effect on the long-term health of the eye, but that lesions and fused cells don't make him comfortable.
He was overall definitely enthusiastic, but I think those two misgivings are definitely something to mention to your eye doctor when discussing the procedure.
I'm afraid you don't remember your TCP/IP correctly- icmp is in no way necessary to open up a tcp connection. It is useful, though, for host unreachable messages from the remote router. But, barring that, ICMP is unecessary for a tcp connection.
Of course you can turn off ping. It's a networking protocol, and like most networking protocols, they can be turned off. Ping is not handled at the hardware level. That's why icmp can be filtered via ipfw in FreeBSD. You can even specify what types of icmp messages to permit. And using sysctl, you can disable certain icmp behaviors in the kernel itself. So, its not handled at the hardware level.
Matt Zito Who, thanks to his job, now knows more about FreeBSD than Linux.
Slashdot Mirror
83! It's easy to remember, simple, and comes default in 100% of linux distributions.
Of course, there are those who are type 82 bigots. I can see how that's important, but with RAM prices so low these days....
Matt
I work for the company managing the technical backend for .pro, and while I understand you're saying this at least partially tongue-in-cheek, I need to correct your misconception. .Pro is for licensed professionals onle - at the time of veritifcation, license numbers, etc. will be submitted. There will be subdomains of .law.pro, .med.pro, etc.
Now, whether that's useful to anyone, I don't know. But no finnair.pro, I'm afraid.
Thanks,
Matt
Note: I work for a domain registrar, but I don't speak for one.
.ORG was never intended to be for just non-profit organizations, and has never been enforced as such, except perhaps at the registrar level by ignorant registrars. According to RFC 1591 by Postel himself:
.ws is a country-code TLD - in this case its for Western Samoa. The concept of .ws being for a "website" is purely marketing.
ORG - This domain is intended as the miscellaneous TLD for organizations that didn't fit anywhere else. Some non-government organizations may fit here.
See - its miscellaneous. In addition,
As far as polluting the namespace - the DNS system is designed to support a large number of TLDs. The restraints that need to be placed are policy ones, not technical ones.
Thanks,
Matt
Note: Although I work for a domain registrar, I don't speak for one.
You obviously haven't read my post closely - ICANN doesn't want to limit the number of TLDs, but it would be foolhardy to open the floodgates without at least giving a testbed to examine the implications of adding TLDs. As for them taking years to do this, it would be just as foolhardy to fundamentally change the structure of the internet without examination and testing of changes they'd like to make - incremental movement rather than revolution.
Thanks,
Have you dumped the root zone lately? The root zone is only about 60k in size. Each TLD only requires about six records - a couple of NS records for the TLD, plus the glue A records. the GTLDS(.com, .net, and .org) are on the gtld servers, and they're the ones with the 2.5 gig zone file. The load on the root servers will increase a negligible amount if 200 new TLDs were added tomorrow.
Thanks,
People really don't understand how the process that the Internet is going through with regards to new TLD additions. This is a TESTBED phase - the whole concept being that they try adding several new tlds with varied scope and purposes to see what the implications and results are. This is the same process that was done when Network Solutions was the only domain name registrar. 5 testbed registrars were chosen to see what would happen when the market was opened to competition - one year later, the market was opened up as a whole and today there are several hundred accredited registrars, and one of the top 3 registrars (Tucows/OpenSRS) was not even one of the testbed registrars.
.com. Come and get 'em
The reason the testbed phase is important is because of exactly the reasons people are complaining about the new TLDs. Now that people have noticed that there's a problem with trademark holder verification, perhaps when the process is opened up (which ICANN has said it will, assuming the testbed phase works) that will be remedied. You have to look at this as a feasibility experiment. Look at the new TLDs:
-.pro - restricted use, but unique in that it provides identity/professional proof of id (for lawyers, etc)
-.info - unrestricted tld, just like
-.biz - for businesses only, iirc. Semi-restricted TLD
-.museum - very specific restricted TLD
-.coop - for non-profits, etc.
This is a textbook example of what should do for a feasibility study - select examples of each type and put them into production. See what happens. Make note of what works and what doesn't and use that to formulate an overall policy.
The idea that ICANN is somehow for limiting the number of TLDs is ludicrous - everyone, from the internet populace at large up to domain name registrars, want new top-level domain names. Everyone would win. But ICANN cannot simply open up the field without understanding and learning about what the implications are. People are looking almost entirely at the technical issues with adding new TLDs, while completely ignoring policy and procedural issues. Issues like false trademark submissions only prove how necessary this process is.
Thanks,
The other issue with Fibre Channel is that its simply SCSI-III encapsulated in Fibre Channel Packets, so there's a lot of overhead.....which makes Fibre not quite the magic bullet that people say. A fibre channel link is only good for maybe 70 MB/sec. because of signaling and overhead. But, on the plus side, You can take your fibre channel loop, stick your servers at the ends, and run IP over Fibre Channel - you just need to make sure your cards support is. I don't think any do under linux, but under Sun most do.
Thanks,
Matthew J Zito, CCNA
Yes, yes, yes.....the cost is that low when you look at the simple cost of the drives. Then you need to add up all the other costs:
-Salaries of people to run it: 1 million/year (20 people at an average of 50,000/year each, which is low)
-Database software: Assuming you want to run oracle, that'll cost you a variable amount depending on the number of cpus you want to run it on - say a 16-cpu box...... 400,000
-Storage system - I know the cost of drives are 50,000, but you need the software, the support infrastructure - 10 terabytes of storage doesn't just show up in a box. You need something like an EMC. Let's say it is an EMC, and you get a good deal - That'll be about 8 million dollars (you may think I'm kidding or exaggerating, but I'm not)
-Database server- this database won't run on a linux box - you're talking something like Sun. I don't know IBM or HP boxes, so I'll price out a Sun - a 16-cpu, 16gb of ram sun e6500 will run you about 600,000 and you need two of them (at LEAST) - 1.2 million
-HA software - Stuff like Veritas, your backup software, software raid, cluster server, etc. - 400,000
Now we're up to 11,000,000 dollars, and that's initial costs, not counting co-lo space, front-end application servers, backup servers, bandwidth, office space, benefits, management, etc. etc. etc. Of course, if you're a pre-existing company, you may be able to leverage some of your existing infrastructure to support this, but no matter what, there's no way this will cost 50k.
Thanks,
Matt
Matthew J Zito, CCNA
At my work, we run two large databases, one that's about 95 gigs and the other 160 gigs on linux. Now, we run our production db on solaris, but the data warehouse and the ticketing db are on linux with oracle. We've had great results with linux for the most part - the biggest problem is that the documentation by Oracle is not as good for linux as it is with Solaris, and its harder to find DBAs with Linux experience.
I can't vouch for Sybase's stability under linux, but Oracle will do you just fine. Get a dual or quad-cpu box, depending on how much data you need to do, and 2 gigs of RAM either way.
Matt
Matthew J Zito, CCNA
Well, when you're at the 30 gigabyte size, your options as far as remote replication are a little limited. For example, its not enough to warrant the sort of large-scale storage array that an EMC Symmetrix would offer that comes with built-in remote replication (SRDF).
What you could use is Veritas Volume Replicator. It runs as a service/daemon on your box and mirrors every write over IP to another box. It can be configured to do it synchronously (the I/O blocks until the remote I/O completes) or asynchronously (higher performance because there's no delay, but you run the risk of data loss when the db server goes down).
Unfortunately, Veritas VR is not available under linux - I think you said it wasn't under linux anyway, but a lot of people are offering linux solutions.
Also, given that your database is only 30 gigabytes, do you actually do a lot of writes? Realistically, if you only do a couple of hundred inserts an hour, you could just, every hour, manually insert the changed records into the remote db. Heck, do it every 5 minutes. I'm not familiar with Sybase, but on Oracle, you can just run the redo logs on the remote data center. That's going to be the cheapest option, and the most linux compatible.
Anyway, if this is really enterprise-level, spring for Veritas - their stuff is expensive but really good.
Cheers,
Matt
Matthew J Zito, CCNA
My GOD! I see it all so clearly now! Roblimo HIMSELF has been assimilated! Who's next? CmdrTaco? OR..dare I say it..JON KATZ!
We all need to send shock troops right away to the Geek compound.
In all seriousness, though, this article raises issues that I think most of the Slashdot readership are aware of, like the fact that although Microsoft may have questionable business practices, their coders are real human beings, many of whom are just trying to have a normal programmer existence. Its only the trolls, flamers, and l33t ones that somehow delude themselves (or act like they do) into believing that Microsoft is roughly equivalent to the 3rd circle of Hell.
Now, something else I find interesting that isn't brought up is whether these trolls and flamers are hurting the linux community as a whole. We've seen in the past where people have posted negative linux articles on places like ZDNet and been flamed up and down. Next time they have to write an article, do you think they'll pick something about linux, knowing that they could spend a fair chunk of the rest of their day just sorting out the flames? Or will they just write an article about NT, or Sun's newest upgrade to their Starfire system ?
This doesn't seem to happen as much (at all, really) with Sun articles, or Oracle articles, or even Windows articles. I think that's because the grass-roots, "us-against-them" mentality that has spurred a lot of really great linux development, press, and most importantly users has also created a jihad in the minds of some people who see anything anti-linux as a death knell for linux. I think linux has been raised to a certain public awareness, though, where we can take some bad articles, get the useful information out of them, and move on. If its FUD, fine. Everyone spreads FUD about everyone else, which doesn't make it okay, but DOES mean that its not some microsoft conspiracy.
Which, after a long tangent, brings me back to the original post.
Matthew J Zito, CCNA
I work for one of the major registrars - NSOL simply runs the SRS (Shared Registry System). Your domain contract is with your registrar. All the registrars sign a contract with ICANN regarding domain disputes, etc.
Matthew J Zito, CCNA
Well, but the point is that the prevailing theory among genealogists was that any given surname originated from a number of people - this suggests that that may not be the case.
Matt
Matthew J Zito, CCNA
An alarm pair is (IIRC) a lower-bandwidth P-t-P circuit that is routed differently through the Telco network. Wish I remember more specifics about it, but I've never ordered one myself. There's no QoS guarantee for circuit noise, etc. Caveat Emptor.
Regarding the actual implementation of this, there was a discussion on this topic on both the ISP-EQUIPMENT and Cisco GroupStudy mailing lists. The general consensus was this:
1) Make sure the distances from the PoP to the CO is less than 12,000 feet for the best chance.
2) Not all Bells will allow you to get these unloaded copper pairs, and those that do will often not give you the highest quality circuits.
I've never done it myself, but there seems to be a fair amount of success in the matter in the marketplace.
Matt
Matthew J Zito, CCNA
Of course, that was in 1997.
Since then, everyone beat them into the ground. This is an almost-verbatim transcript of a conversation between me and a BSDI engineer about 6 months ago:
Me: So, do you support the IBM Netfinitys?
Engineer: No.
M: Do you support Fibre Channel?
E: Maybe by the third quarter of 2000.
M: Do you support Gigabit Ethernet?
E: No.
M: Do you support clustering or shared arrays?
E: No.
(3 weeks later, I call the guy back)
M: Hey, we've fired up BSDI on a netfinity, but the 10/100 ethernet card in it will only run at 10 megabits. Its an AMD PcNet 100.
E: Oh, that driver only supports 10 megabits.
M: Any timeframe on when that'll be fixed?
E: I couldn't say.
Needless to say, we kept freebsd.
Matthew J Zito, CCNA
5541139af5ed9ac22e855c30e6927f54
Again, if anyone gets a better source, send it to me!
Matthew J Zito, CCNA
I find it very interesting that some sort of serious legal authority is getting involved in this situation - I was under the impression that in the U.S. it was a civil suit. Is it a civil case in Norway also? Who are these people? Can someone fill that in?
This was a wake-up call for me. I was supportive of the EFF and everyone else involved in the suit in California, but I figured it would be one of those things where the case went to court, the judge laid the smack-down (so to speak) on the plaintiff, and we all went on our merry way. But the fact that an indictment was returned against this gentleman shows that not only is the DVD CCA intent on making a serious effort to put a stop to legal and legitimate reverse engineering but that there is a severe possibility that WE MIGHT LOSE.
I went and read the response by the DVD CCA- their argument is since the code has the master key in it, it MUST have been obtained illegally. This is a circular argument. But if they are able to convince a judge that this is true, this could signal an end to the idea of black-box reverse engineering.
How can we prevent this from happening? I'm starting by putting a copy of the DeCSS code up on my personal web server: http://128.122.106.158/decss.txt This is the only code I have - if someone wants to email me something more complete, I'll put that up. Email me at matthewzito@yahoo.com.
Everyone should put this code on their site. If enough people put this code on a website somewhere, the DVD CCA can't sue/arrest/harrass everyone. It's an old, hokey protest tactic, but it works.
Next, contact anyone and everyone in power. Call newspapers, politicians, and tell anyone and everyone who will listen. If they won't listen, tell them multiple times. Be polite, but be firm.
Write letters to the editor. Here, the idea is to make sure that everyone is aware of the issue, and more importantly, is aware of our side of the issue. Make sure that if someone reads an article that supports the DVD CCA, they have already read or heard something from someone sympathetic to our cause.
Donate money to the EFF, and any legal defense fund that is created for this gentleman. It doesn't have to be a lot, but anything you can give can help insure that we get a victory in the courts.
Anyone interested in putting together a little fund to run an ad in a major US newspaper like the NY times with the DeCSS code in it? That's more of a farfetched idea than most of the others in this post, but its a beautiful idea nonetheless.
Basically, the only things we can do are practice social disobedience (even in such a minor form as keeping our own public copies of the decss code), support those who are fighting the legal battles in the courts, and educate, educate, educate. Even if we lose some court battles, if the public in general is aware of the issues at stake, that gives us an advantage for future confrontations with companies trying to stop legitimate reverse engineering. Which there will certainly be.
Email me at matthewzito@yahoo.com
Matthew J Zito, CCNA
Matt
Matthew J Zito, CCNA
Well, this is true to a certain extent, but there's some more details that need to be filled in:
Corporations are individuals under the law. In his original post, Mr. Rideau uses the word "companies". There are two general types of companies: Proprietorships and Corporations. Proprietorships are the most common form of companies. The different types of proprietorships are the Sole Proprietorship, the General Partnership, and the Limited Partnership Under a proprietorship, there is no separate legal entity- the owner's possessions and assets are open to legal action. In that situation, the transfer of data from one employee to another constitutes a transfer of the so-licensed software from one individual to another. Therefore, either individual would have the right to re-distribute that software as they saw fit. Even a non-disclosure agreement would not absolve one from this responsibility, since the GPL itself says (emphasis mine):
So, based on this, if one is employed at a company that is a sole proprietorship (which is the most common type of company in the US) and is bound by a NDA not to release or discuss the source code to the public, they are not allowed to redistribute any of the changes back into the company, since that would be distributing the source code in violation of the above conditions.
Now, while a corporation IS its own entity under the law, there is something called "Respondeat Superior", or "the acts of the agent are the acts of the owner" (that information, and the information in general regarding rights of agency courtesy of "Small Business Advisor" pg. 446-449). If an employee of a corporation starts using GPL'ed software, and does so on company time, is the company then bound under the GPL? It depends. If the person in question either has a) a clearly stated authority to speak on behalf of the corporation (i.e. in the articles of incorporation) or b) an implied authority to do so (like you're the head of the software design department and the CTO just said, "We need you to pick a new piece of software to build"), then the corporate entity is bound by the actions of the individual.
Therefore, if a someone within the company who meets neither of the above conditions takes a piece of GPL'ed software, modifies it, and then distributes it to five other employees, s/he is bound by the GPL, NOT the company. OTOH, if someone with authority starts using a piece of software, the company is then the entity bound by the GPL.
Based on that, since Mr. Stallman said that redistribution within a company is not redistribution in terms of the GPL (we have to assume that he means a corporation, since its been demonstrated that transactions within a proprietorship constitute transactions between individuals), it is legally possible to create an organization like the one Mr. Rideau described, where upon entering the corporation you agree to not distribute internal corporate information. Since it is not redistribution according to Mr. Stallman to distribute within a corporation, then section 7 of the GPL with regards to anti-distribution and restriction thereof does not apply, since a new instance of the GPL has not been created. All the employees of the corporations consitute a part of that single entity, so their individual activities within the corporation can be managed.
The upshot of all of this is that Mr. Rideau is right. Acceptance of the GPL by a person in a corporation with the right to do so binds the corporate entity to the GPL. Since internal distribution of the code is allowed, employees can be bound legally not to distribute outside the corporation, since they have not been bound under the GPL - the corporation has.
Imagine a situation where someone takes a piece of GPL code, say Linux, as the most obvious example in this forum. A corporation is formed, and the CEO accepts the GPL. The corporation makes a new, improved version of linux (for the sake of an example, we'll say they rewrite the memory handling code and double speed). Normally, if they wanted to redistribute, they would have to make public their changes. Instead, though, they make the software available by selling small stakes in the company (or simply on a fee basis and making them part of the corporation). So, if you want the software, you have to become part of the corporation. Once you are part of the corporation, you're bound by the rules of the corporation regarding redistribution, not the GPL (since the GPL does not apply to internal corporate redistribution). If the company wanted to release the code to anyone outside of the corporation, then they would have to make it public. But since they're only giving it to other people within the corporation....
Now, I realize that this is a remote situation. But its important that we figure out these weaknesses in the GPL before someone else does and exploits them.
Matthew J Zito, CCNA
I don't buy into the "only gurus should be messing with such mystical things" argument, but at the same time, if someone doesn't configure their software properly, its perfectly acceptable for someone to step in and say, "Because your software is indirectly contributing to a problem, we are going to put a stop to it". Hopefully then the user will step back, find the problem, reconfigure it, and everyone will be better off. The blame isn't on @home for this. Yes, they probably could have been more on the ball about dealing with the customers who were spammers, but as an @home customer, I know a) how many people are running WinGate without really having a strong grasp of the issues involving that and b) how many portscans float by my machine on a day to day basis.
I think the UDP was warranted, and I believe that the response from @home was also. It is people with unsecured proxies that contribute to the problem, not "anyone using a home network". As I said previously, the only ones who will be affected by this scan are those whose proxies are incorrectly configured.
Matthew J Zito, CCNA
I remember that because the freebsd main ftp site got hammered the last time Slashdot reported right away that a FreeBSD release was out, the FreeBSD core team asked slashdot to hold their posts of new releases until the ftp mirrors catch up.
Matt
One of the problems with consumer Internet technologies (i.e. ISDN, SDSL, etc.) is that there is less of a QoS guarantee. In my area, if your ISDN line goes down, Bell Atlantic tries to get a service technician out in a few days, generally more like a week. When you're a small company starting out that could kill you. By contrast, when you have a Frame Relay (also sometimes inaccurately called a Burstable T-1) or T-1 link, Bell will immediately look into a service failure. While they're still not responsive, there is a much higher commitment to service for the higher-end data services. You also get a day of service free when there's an outage that lasts longer than 20 minutes.
.02 about the whole issue RE: Colocation vs. home data line is that colocating will always be cheaper, given two companies of similar quality. That is, colocating with Exodus (excellent NYC colocation provider) is about $1000/month per megabit. By the same token, a T-1 through a high-quality provider like Sprint will run about $1100-$1600/month (depending on the circuit charges). You can get really crappy colocation for as low as $100/month.....just like everything else, you get what you're willing to pay for. Definitely shop around. And ask questions like:
My
1- What hours can I physically work on my server?
2- What kind of technical support do I get?
3- Do they offer remote power management (some colo places give you a web page where you can power cycle your box- its great)?
4- Traffic analysis?
5- Will my server be connected to a hub or a switch? (cheap places tend to hook you up to a cheap hub- this will degrade your network performance and allow people at other colocated servers to sniff your traffic)
Ask these questions and in an old Ask Slashdot there was a thing about co-location providers...go dig that up. There was a lot of really useful stuff in there.
Good luck.
Well, I asked my eye doctor about the surgery about 6 months ago, when I had my last exam. He said I'm an excellent candidate, etc. etc. etc. but then talked about his misgivings about the procedure:
1) He doesn't recommend it until you're 25- he says that before then your eyes are still growing and changing.
2) There have been no long term studies on the after-effects of the surgery. He described how before the surgery, the cornea is a grid of perfect hexagons, and after the surgery, the grid is gone, replaced by fused cells, small lesions, etc. He said that theoretically that should have no effect on the long-term health of the eye, but that lesions and fused cells don't make him comfortable.
He was overall definitely enthusiastic, but I think those two misgivings are definitely something to mention to your eye doctor when discussing the procedure.
Matt Zito
I'm afraid you don't remember your TCP/IP correctly- icmp is in no way necessary to open up a tcp connection. It is useful, though, for host unreachable messages from the remote router. But, barring that, ICMP is unecessary for a tcp connection.
Matt
Of course you can turn off ping. It's a networking protocol, and like most networking protocols, they can be turned off. Ping is not handled at the hardware level. That's why icmp can be filtered via ipfw in FreeBSD. You can even specify what types of icmp messages to permit. And using sysctl, you can disable certain icmp behaviors in the kernel itself. So, its not handled at the hardware level.
Matt Zito
Who, thanks to his job, now knows more about FreeBSD than Linux.
And almost prefers it that way.