"bounc[ing] the stuffing" out of your monitor isn't really the epitome of quantitative analysis now is it?
Let's try this, a little high school physics.
P=E*I
P = is watts E = Voltage I = Current
E=3.6V (the volate of my nokia phone's battery) I=900mAh (the total current avaiable from the battery)
Now,
3.6*900 = 3.2 watts
Indeed, that is the total capcity of my battery, in watts.
So with 3.2 watts from my wee battery lets see how long @ 200W we could sustain that output:
3.2/200 =.016 secs.016 seconds and the battery would be completely dead. Now lets put this against your claim that @ 1% duty cycle this would be reasonable..01 * 180 minutes (total talk time) = 1.8 minutes @ 200W !=.016 Seconds @ 200W
There is a reason that my phone manual says "Transmitter output Up To 600mW"
Also, unless you've missed the memo, the orion project is fiction, as in not real.
When you have something besides unproven guesses and rediculous postulations feel free show the world that you've caught up.
I think you are full of shoe. Prove that a cellphone outputs 200W of "power" - ever. Then go and take the maufacturer to court for exceeding the FCC's output level. If you are going to arge, prove your point instead of more guessing.
The following scenarios are why the MPAA and this corporate IP nonsense needs to be stopped.
Your cable provider sends a notice to all of its customers. It says that due to recent changes in local legislation that only Dell GX150 Desktop machines can be used with their cable internet service. All other equipment is no longer authorized to connect to the internet via the cable service. This then qualifies all non Dell GX150 devices that are communicating with the cable network as "unauthorized"[1]. Depending on how the courts interpret the law - this could make all "unauthorized communications devices"[1] (non-GX150), "unlawful access devices"[1]. A user in possession of an unlawful access device could then be prosecuted under this law for using a system that is not a Dell GX150.
Certainly that is an extreme example though not completely implausible under the current stipulations in the law. It would be useful though to examine a more realistic scenario:
The law is passed in your state legislature - but there are no apparent changes in policy with your service provider (the aforementioned cable provider). In the coming months and years after the law is passed, the MPAA and RIAA find out that there is a "malicious" user illegally "swapping" large amounts of "unlicensed" music in the form of MP3's over the cable provider's network. The RIAA and MPAA approach the cable provider about this user and threaten legal action against the cable company if it does not terminate the user's connectivity. The cable company, would have an easy time under this law finding "unauthorized" devices that are connected to their network,. There are provisions that make it a "crime to "receive... transmit, [or] re-transmit" any communication service without the "express authorization" of the communication service provider."[1] The benevolent provider when pushed to find a reason to terminate the user's connectivity - can easily find one. And then the MPAA and RIAA can do as they wish with you.
The real problem with this law is the broad terminology used and the fact that it can be so easily used to deny ANYONE communications services. The term "communications device" can mean anything. The dictionary definition of the term "device" could be argued to include network protocols as a "device for communicating", or even software that you have written that uses a network provided by someone else. The ambiguous terms defined in most of the versions of this law such as "communications devices" are dangerous for end users because they leave so much open for definition and interpretation. They are designed to target end users, so that the creators and proponents of this bill may engage users in direct or indirect legal action. An area where individuals are not well equipped to defend themselves.
These scenarios listed here are just the beginning - the implications of this facet of the law are not limited to just the Music industry or consumers. They apply to communications everywhere.
The simple matter of it is - this law is an "Exploit". The MPAA and RIAA are implementing the law and variations of it in the state legislatures because they are lower profile areas where legislators are less cognoscente of the implications of this bill. As a result this law will/has already grant the RIAA/MPAA the ability to influence those who do not have significant legal clout to do as they would see fit. Though, the MPAA and RIAA are only the most immediate concerns - the broad and general definitions laid out in this law can be used by any entity to deny communications services to an individual or organization. Fred von Lohmann speaks of transferring law enforcement from public to private hands, this is exactly that. Though, law enforcement is the exclusive domain of the Government. This law bypasses our legal system and allows other parties to prosecute under the guise of a "law". The only reason this can happen is that the state legislators have and insufficient understanding of the implications of s
Dude - you missed it. The law allows the provider to ban ANY device connected to their networks. It doesn't matter if it is TiVo or TV the law is so broad and sweepeing that it takes all "communications devices" into account - that is to say - anything that you use to communicate with, anything that receives or transmits a coded transmission basically.
So unfotunately when your cable provider gets sued by the MPAA to dis allow the use of TiVO then you will be on the receiveing end of this law.
Clearly you have not thought about the implications of such a proposition.
First, you need a cable drop for every access point that you establish. This reduces the cost per unit, but doesn't really eliminate the problem.
Second, each access point has a limited range some people will get a better signal than others. Realistically with a 54Mb access point you'll be able to support ~ 12 Users @ 200K/s the max of a T-1 or 1.5 Mbit DSL. The Pipe to the condos should provide at least this level of service at a price below market value for this project to even be justified. Also, the number of access points could be few or many depending on the building types. What this translates into is a lot of up front testing before a project plan is created for doing the cable pulls.
Thrid, wireless on desktops isn't really a good deal unless you have a Mac. PC's have pretty whack PCI Wireless cards and the technology is much more cumbersome from an enduser and administrative point of view. I.E. a lot more can go wrong between the desktop and the AP.
Fourth, security. It can be almost as secure as a wired network but the effort to accomplish this is much higher. Time = Money.
All in all wireless is the last option the person should choose. In a situation such as this - it is a last resort for people who can't do a LAN the way it should be done.
I propose that we build a database and client application in which people can connect and download the offensive unicast ads thereby making their ad formats untrackable. The full ads won't be partiucularly useful if the activity and impressions can't be accurately tracked (becuase of all of the downloads by people participating). In addition the "number of impressions" that are served will be inaccurate and advertisers campaigns will not run as long.
It looks nice for sure - but the specs... are lacking. The reason they can put the VIA proc. in the case w/no fan is simply that it's a "smaller engine" that generates less heat. They haven't done anything "revolutionary" here. The have just put a low power processor in a closed case with enough sinks on the side to disapate internal heat. Apart from surfing, office, music and possibly watching a DVD the proc in the machine isn't good for much. As for our enthusatic posters question "Is this finally the step to having a true PC in every living room? HTPC here we come!" No - it isn't. This will nto be the path for faster processors, they generate too much heat.
They simply can't be compared. The reason for this statement is that AAC "supports" or "is encumberd by" (depending on what camp you are in) digital rights management. While they are all formats for redistributing music, OGG is not an option when trying to negotiate with record companies who need some assurance that their music won't be redistributed.
A better comparison would be WMA vs. AAC and OGG vs. MP3.
Technically you have to pay for windows. This in itself makes it more diffcult to obtain. There are no restrictions on the distribution of BSD. Perhaps it is a little paranoid. I'm not sure that we should be "giving" our enemies tools that help them secure their communications infrustructures.
Yeah - that is a problem perhaps... what if the point of access is in a different state:-) Basically this law is just Krap. I'd like to see a national consumer rights act through congress that would a put an end to this garbage.
The "Super DCMA" is without question a disturbing "revelation" on behalf of the entertainment industry. It shows the continued distrust of business of the consumer, and the desire to engage in anti-competitive behavior by permitting the revocation basic consumer rights.
If taken to an extreme it is possible that a prominent cable company could go so far as to say that you may only use dell computers with their service and that not adhereing to this is a violation of the law. Dell could become a "monoply" by entering into a "deal" with communications providers when in fact that deal - violates anti-trust laws. It certainly seems "bad". Though this sort of legislation isn't enforceable.
Take NAT for example. Many have been saying that this bodes ill for network address translation. I submit to you this: NAT is most commonly used in a Local Area Network environment. In large companies it is used over a WAN. In either case, it can be argued that the traffic and the origin of the traffic is well known. Each machine has an IP address and whether it is "translated" or not, the communication orginates from a well known PHYSICAL location. The physical origin of the traffic is well known, therefore the law isn't being violated. Clearly - it is harder to defend against this law if the location and "origin" of the offending machine in question is not well known...
What is interesting is how they are using the state legislatures (less visible and more malleable) to enact this sort of law. If it was really legit, then why not go the federal route?
Finally, this sort of law is a golden opporunity for a service provider to provider service that is competely unencumbered by the provisions of this "Uber DCMA". It's only valid if the "commmunications" service provider decides to enforce it.
d00d - this is a remote display. It's meant to augment your exising system with a touch display that is mobile. As a display it's problably not that great - how many pixels can an XScale Push if an XScale could push pixles?
It's first generation and it shows - but I think that we will all one day have some thing like this. Honestly why sit in front of or stationed near a computer when all you really want or need is the display? And how much typing does it take to read/. for any real work, you can get back to the keyboard.
Make is smaller faster cheaper, oh wait that's NASA and we'll be on to something!
I can save money both on hardware costs and labor costs. I can also build better and more reliable, less expensive web applications using PHP instead of Java.
Perhaps in your world wasting time and money is OK. But not in mine.
computers aren't really cars as far as warranty's go. it's hard to justify a 2 year difference in warranty period. a similar server to the dell from ibm has a 3yr. warranty also. Sun parts cost more too.
To be honest I beleived this for a while too, until I read an article about the 2.6 kernel. Premptive multitasking? Welcome to 1995. Separate queues for each block I/O device, isn't this common sense?
is this really needed for the applications that we are talking about? solaris IS more advanced than linux yes I haven't ever refuted that, but linux is more than adequite and costs less to own than solaris.
I'll give you that Gigabit Ethernet isn't totally necessary. It was one of the weakest of the arguments. thought the chipset provided by dell allows for load balancing and failover in hardware.
hardware raid IS important, very important, these machines are typically @ the front lines of handling sometimes unpredictable loads. the reason FOR hardware raid is primarily to augment the processor in pushing bytes to the disk. sure linux and M$ both do software RAID to and it's no big deal. but for the price that dell offers it - there is no reason to offer it at well. in my mind reliability is supported by redundancy. ALSO you have to agree that a 1 year warranty is inexcuseable. Of the past sun machines that i've used two of the three broke. one needed a mother board replaced in the first few months after buying it and the other's ethernet interface died after being shipped somewhere (bizzare). My feeling is generally that a machines warranty should be 2/3rds of it's useful lifespan.
i'll find you benchmarks...
just becuase thousands of businesses use java (i do) doesn't mean that it's good. but that was a bit of an exageration.
My point with the 3 for every two was primarily that you get 3 servers that are more expandable for the price of two that are not... i'm not sure that they are all that competitve primarily because the v120 isn't really playing in the same league as the dell, basic things like second processor and the third drive bay and the redundant PS.
again, i think sun has a place - but it's not in the emerging linux market. and to my point last post we will find that companies are going to move away from sun hardware in the future. give it another 3 to 4 years. only time will tell - chalk one up to philosophy.
I've left the rest of the computing world out of the discussion becuase I think that the markets that we are talking about, those are web servers, network utility machines i.e. DNS, DHCP, Firewall and to some extent File Servers don't have any other real competitors but M$.
AIX, IRIX are the highend OS's that Sun needs to go compete against - Sun has the hardware line and the OS to do so. They continue to waste their money with lowend servers that will never reach a price point that is competitve against AMD and Intel.
With the advent of 64-Bit chips from AMD and Intel on the horizon, clock cycle for clock cycle AMD and Intel hardware will have the lowend single proc and dual proc sun boxes out matched. as far as remote management goes see the price comparison post... a full lights out remote management card from dell is $399. That's all that is needed to be competitve. Beyond simply clock cycles the Intel architecture is about to undergo another evolution. the new bus architecutures that will support switched bandwidths greater than 2GB/sec will lead to massive disk and network throughput with the help of SATA and 10 Gigabit ENet.
the interchangeability of hardware is important from an IT perspective. having parts laying around, or being able to repurpose a boxe's task is important most netoworks are still mixed usally containing some form of windows - the linux servers that get decomissioned can still be used for other tasks that require windows. the same is not true for sun machines.
In the end I'm not saying that linux categorically eliminates solaris from the world, that is not the case at all. I'm trying to make the point that Sun is playing in a world in which they do not have the ability to compete, one that linux and intel have licked. There are other markets that will have a need for some time for the high end n-way boxes that sun produces but not the areas in which linux is being adopted. Linux is not the one stop shop answer to all the licensing woes of the personal computing and IT world; but sun's "brand" no longer carries the same singular sense of reliability and depend that it once did - becuase there are alternatives. Change is slow in the IT world - with the dot com boom gone *boom* - saving money is cool, and budgets are slim. these all point to Sun >rapidly< rethinking their current direction.
In two to three years when the existing dot com then "legacy" systems come up for upgrade - intel and amd hardware will be very attractive from a price/perfomance standpoint AND they will have a *nix esq OS that is available for deploying all of their Linux ported applications. The choice, will be clear.
Sun v120 (There is no v150) 1U SINGLE PROCESSOR CAPABLE @ UltraSPARC IIi 650 Mhz 1 Gig RAM 2x 36 gig SCSI 10K Drives 2x10/100 On Board-Nic 1x1 Gigabit Ethernet ( added to compete with dell svr.) UltraSCSI II (No Hardware R aid) cd-rom No redundant power supply 1 yr. wrty. Price as configured: $5,757
DELL 1650 1U DUAL Processor Capable w/ ONE processor @ P4 -1.24Mhz 1GB RAM 2 x 10K 36 GB Drives Hardware RAID - UltraSCSI II On-Board 2x 1 Gig E-net On-Board cd-rom redundant supply available 3 YR. Warranty Price as configured: $3,640
basically for every two sun servers - you could get 3 dell servers. they cost less to upgrade too.
under most circumstances i'd recommend that drives and supplies be redundant - it doesn't hurt and can only help. Sun charged 1K for the extra drive @ purchase time. Most notably sun lacked the one gigabit on-board ethernet.
A server's sole purpose is to push bytes and process. slower nic's = slower performance. The lack of mutli proc on sun's part isn't surprising - you have to pay for expandability. Also of note is the lack of redundancy i.e. Hardware RAID support and redundant power supplies.
The v120 is their "entry level server" that is clearly lacking in expandability, and reliability. As stated before - sun hardware at the web server level cannot compete. just as sun does a shoddy job with their software - they do the same with their hardware - and they won't even warranty their hardware beyond a year. Repairs aren't cheap.
As for maintainance - starting adim salary for linux is lower than that of sun. also, because linux is freely available - the likleyhood that a candidate has spent more time working with linux boxes in and out of a work environment is higher.
there was a study posted on/. a few months ago about the cost of ownership for windows, linux and solaris. solaris wasn't even in the running. on going costs were vastly higher than the other two platforms.
Sun has demonstrated a particular ineptitude in software development. For example Java has become a rather messy collection of API's. Some might call it "robust" however Sun constantly defines new API's touting them as "innovations". However there is no leadership in execution. Why is it that developers are forced to rewrite sun written code to enchnance performance and make their applications reliable? The truth is that Sun is very fond of "reference" implentations. I.E. they are half assed, untested and unreliable. To get really good perfomance other have to start from the ground up. Blackdown, IBM's 1.3 JVM comes to mind, Caucho's XSL Parser and there are others. Sun's leadership with repspect to software is really none at all. Setting up a portal and creating "standards" through API's isn't really adequite.
The restriction that I was refering to was primarily hardware. Intel hardware, the most common and definately most compelling hardware platform to deploy linux on does not scale particularly well. If a machine was needed that required 32 processors IBM or Sun would likely be the only two choices. It will be a long time before linux will be able to play in this arena. There will always be a need for these machines.
Slashdot Mirror
He IS an ass clown!
you my friend need a clue stick.
.016 secs .016 seconds and the battery would be completely dead. Now lets put this against your claim that @ 1% duty cycle this would be reasonable. .01 * 180 minutes (total talk time) = 1.8 minutes @ 200W != .016 Seconds @ 200W
"bounc[ing] the stuffing" out of your monitor isn't really the epitome of quantitative analysis now is it?
Let's try this, a little high school physics.
P=E*I
P = is watts
E = Voltage
I = Current
E=3.6V (the volate of my nokia phone's battery)
I=900mAh (the total current avaiable from the battery)
Now,
3.6*900 = 3.2 watts
Indeed, that is the total capcity of my battery, in watts.
So with 3.2 watts from my wee battery lets see how long @ 200W we could sustain that output:
3.2/200 =
There is a reason that my phone manual says "Transmitter output Up To 600mW"
Also, unless you've missed the memo, the orion project is fiction, as in not real.
When you have something besides unproven guesses and rediculous postulations feel free show the world that you've caught up.
I think you are full of shoe. Prove that a cellphone outputs 200W of "power" - ever. Then go and take the maufacturer to court for exceeding the FCC's output level. If you are going to arge, prove your point instead of more guessing.
The following scenarios are why the MPAA and this corporate IP nonsense needs to be stopped.
... transmit, [or] re-transmit" any communication service without the "express authorization" of the communication service provider."[1] The benevolent provider when pushed to find a reason to terminate the user's connectivity - can easily find one. And then the MPAA and RIAA can do as they wish with you.
Your cable provider sends a notice to all of its customers. It says that due to recent changes in local legislation that only Dell GX150 Desktop machines can be used with their cable internet service. All other equipment is no longer authorized to connect to the internet via the cable service. This then qualifies all non Dell GX150 devices that are communicating with the cable network as "unauthorized"[1]. Depending on how the courts interpret the law - this could make all "unauthorized communications devices"[1] (non-GX150), "unlawful access devices"[1]. A user in possession of an unlawful access device could then be prosecuted under this law for using a system that is not a Dell GX150.
Certainly that is an extreme example though not completely implausible under the current stipulations in the law. It would be useful though to examine a more realistic scenario:
The law is passed in your state legislature - but there are no apparent changes in policy with your service provider (the aforementioned cable provider). In the coming months and years after the law is passed, the MPAA and RIAA find out that there is a "malicious" user illegally "swapping" large amounts of "unlicensed" music in the form of MP3's over the cable provider's network. The RIAA and MPAA approach the cable provider about this user and threaten legal action against the cable company if it does not terminate the user's connectivity. The cable company, would have an easy time under this law finding "unauthorized" devices that are connected to their network,. There are provisions that make it a "crime to "receive
The real problem with this law is the broad terminology used and the fact that it can be so easily used to deny ANYONE communications services. The term "communications device" can mean anything. The dictionary definition of the term "device" could be argued to include network protocols as a "device for communicating", or even software that you have written that uses a network provided by someone else. The ambiguous terms defined in most of the versions of this law such as "communications devices" are dangerous for end users because they leave so much open for definition and interpretation. They are designed to target end users, so that the creators and proponents of this bill may engage users in direct or indirect legal action. An area where individuals are not well equipped to defend themselves.
These scenarios listed here are just the beginning - the implications of this facet of the law are not limited to just the Music industry or consumers. They apply to communications everywhere.
The simple matter of it is - this law is an "Exploit". The MPAA and RIAA are implementing the law and variations of it in the state legislatures because they are lower profile areas where legislators are less cognoscente of the implications of this bill. As a result this law will/has already grant the RIAA/MPAA the ability to influence those who do not have significant legal clout to do as they would see fit. Though, the MPAA and RIAA are only the most immediate concerns - the broad and general definitions laid out in this law can be used by any entity to deny communications services to an individual or organization. Fred von Lohmann speaks of transferring law enforcement from public to private hands, this is exactly that. Though, law enforcement is the exclusive domain of the Government. This law bypasses our legal system and allows other parties to prosecute under the guise of a "law". The only reason this can happen is that the state legislators have and insufficient understanding of the implications of s
Dude - you missed it. The law allows the provider to ban ANY device connected to their networks. It doesn't matter if it is TiVo or TV the law is so broad and sweepeing that it takes all "communications devices" into account - that is to say - anything that you use to communicate with, anything that receives or transmits a coded transmission basically.
So unfotunately when your cable provider gets sued by the MPAA to dis allow the use of TiVO then you will be on the receiveing end of this law.
DUUUUUDE a compass is designed to pick up electromagnetic fields... OF COURSE IT'S UNSHIELDED.
Yes, and a compass meausures the magnetic field of the earth. For something like the 100mW is a lot....
Clearly you have not thought about the implications of such a proposition.
First, you need a cable drop for every access point that you establish. This reduces the cost per unit, but doesn't really eliminate the problem.
Second, each access point has a limited range some people will get a better signal than others. Realistically with a 54Mb access point you'll be able to support ~ 12 Users @ 200K/s the max of a T-1 or 1.5 Mbit DSL. The Pipe to the condos should provide at least this level of service at a price below market value for this project to even be justified. Also, the number of access points could be few or many depending on the building types. What this translates into is a lot of up front testing before a project plan is created for doing the cable pulls.
Thrid, wireless on desktops isn't really a good deal unless you have a Mac. PC's have pretty whack PCI Wireless cards and the technology is much more cumbersome from an enduser and administrative point of view. I.E. a lot more can go wrong between the desktop and the AP.
Fourth, security. It can be almost as secure as a wired network but the effort to accomplish this is much higher. Time = Money.
All in all wireless is the last option the person should choose. In a situation such as this - it is a last resort for people who can't do a LAN the way it should be done.
I propose that we build a database and client application in which people can connect and download the offensive unicast ads thereby making their ad formats untrackable. The full ads won't be partiucularly useful if the activity and impressions can't be accurately tracked (becuase of all of the downloads by people participating). In addition the "number of impressions" that are served will be inaccurate and advertisers campaigns will not run as long.
It looks nice for sure - but the specs... are lacking. The reason they can put the VIA proc. in the case w/no fan is simply that it's a "smaller engine" that generates less heat. They haven't done anything "revolutionary" here. The have just put a low power processor in a closed case with enough sinks on the side to disapate internal heat. Apart from surfing, office, music and possibly watching a DVD the proc in the machine isn't good for much. As for our enthusatic posters question "Is this finally the step to having a true PC in every living room? HTPC here we come!" No - it isn't. This will nto be the path for faster processors, they generate too much heat.
They simply can't be compared. The reason for this statement is that AAC "supports" or "is encumberd by" (depending on what camp you are in) digital rights management. While they are all formats for redistributing music, OGG is not an option when trying to negotiate with record companies who need some assurance that their music won't be redistributed.
A better comparison would be WMA vs. AAC and OGG vs. MP3.
HAHAHAHAHHAHAAH - /.'d
Technically you have to pay for windows. This in itself makes it more diffcult to obtain. There are no restrictions on the distribution of BSD. Perhaps it is a little paranoid. I'm not sure that we should be "giving" our enemies tools that help them secure their communications infrustructures.
Yeah - that is a problem perhaps... what if the point of access is in a different state :-) Basically this law is just Krap. I'd like to see a national consumer rights act through congress that would a put an end to this garbage.
The "Super DCMA" is without question a disturbing "revelation" on behalf of the entertainment industry. It shows the continued distrust of business of the consumer, and the desire to engage in anti-competitive behavior by permitting the revocation basic consumer rights.
If taken to an extreme it is possible that a prominent cable company could go so far as to say that you may only use dell computers with their service and that not adhereing to this is a violation of the law. Dell could become a "monoply" by entering into a "deal" with communications providers when in fact that deal - violates anti-trust laws.
It certainly seems "bad". Though this sort of legislation isn't enforceable.
Take NAT for example. Many have been saying that this bodes ill for network address translation. I submit to you this: NAT is most commonly used in a Local Area Network environment. In large companies it is used over a WAN. In either case, it can be argued that the traffic and the origin of the traffic is well known. Each machine has an IP address and whether it is "translated" or not, the communication orginates from a well known PHYSICAL location. The physical origin of the traffic is well known, therefore the law isn't being violated. Clearly - it is harder to defend against this law if the location and "origin" of the offending machine in question is not well known...
What is interesting is how they are using the state legislatures (less visible and more malleable) to enact this sort of law. If it was really legit, then why not go the federal route?
Finally, this sort of law is a golden opporunity for a service provider to provider service that is competely unencumbered by the provisions of this "Uber DCMA". It's only valid if the "commmunications" service provider decides to enforce it.
My 2 pestas.
are you really a teenager? you write awfully well for a lil' chillin.
d00d - this is a remote display. It's meant to augment your exising system with a touch display that is mobile. As a display it's problably not that great - how many pixels can an XScale Push if an XScale could push pixles?
/. for any real work, you can get back to the keyboard.
It's first generation and it shows - but I think that we will all one day have some thing like this. Honestly why sit in front of or stationed near a computer when all you really want or need is the display? And how much typing does it take to read
Make is smaller faster cheaper, oh wait that's NASA and we'll be on to something!
I can save money both on hardware costs and labor costs. I can also build better and more reliable, less expensive web applications using PHP instead of Java.
Perhaps in your world wasting time and money is OK. But not in mine.
computers aren't really cars as far as warranty's go. it's hard to justify a 2 year difference in warranty period. a similar server to the dell from ibm has a 3yr. warranty also. Sun parts cost more too.
To be honest I beleived this for a while too, until I read an article about the 2.6 kernel. Premptive multitasking? Welcome to 1995. Separate queues for each block I/O device, isn't this common sense?
is this really needed for the applications that we are talking about? solaris IS more advanced than linux yes I haven't ever refuted that, but linux is more than adequite and costs less to own than solaris.
competitive, yes. although have a look at the warranty... to get a 3 yr wrty it's an extra 3K... what gives?
I'll give you that Gigabit Ethernet isn't totally necessary. It was one of the weakest of the arguments. thought the chipset provided by dell allows for load balancing and failover in hardware.
hardware raid IS important, very important, these machines are typically @ the front lines of handling sometimes unpredictable loads. the reason FOR hardware raid is primarily to augment the processor in pushing bytes to the disk. sure linux and M$ both do software RAID to and it's no big deal. but for the price that dell offers it - there is no reason to offer it at well. in my mind reliability is supported by redundancy. ALSO you have to agree that a 1 year warranty is inexcuseable. Of the past sun machines that i've used two of the three broke. one needed a mother board replaced in the first few months after buying it and the other's ethernet interface died after being shipped somewhere (bizzare). My feeling is generally that a machines warranty should be 2/3rds of it's useful lifespan.
i'll find you benchmarks...
just becuase thousands of businesses use java (i do) doesn't mean that it's good. but that was a bit of an exageration.
My point with the 3 for every two was primarily that you get 3 servers that are more expandable for the price of two that are not... i'm not sure that they are all that competitve primarily because the v120 isn't really playing in the same league as the dell, basic things like second processor and the third drive bay and the redundant PS.
again, i think sun has a place - but it's not in the emerging linux market. and to my point last post we will find that companies are going to move away from sun hardware in the future. give it another 3 to 4 years. only time will tell - chalk one up to philosophy.
I've left the rest of the computing world out of the discussion becuase I think that the markets that we are talking about, those are web servers, network utility machines i.e. DNS, DHCP, Firewall and to some extent File Servers don't have any other real competitors but M$.
AIX, IRIX are the highend OS's that Sun needs to go compete against - Sun has the hardware line and the OS to do so. They continue to waste their money with lowend servers that will never reach a price point that is competitve against AMD and Intel.
With the advent of 64-Bit chips from AMD and Intel on the horizon, clock cycle for clock cycle AMD and Intel hardware will have the lowend single proc and dual proc sun boxes out matched. as far as remote management goes see the price comparison post... a full lights out remote management card from dell is $399. That's all that is needed to be competitve. Beyond simply clock cycles the Intel architecture is about to undergo another evolution. the new bus architecutures that will support switched bandwidths greater than 2GB/sec will lead to massive disk and network throughput with the help of SATA and 10 Gigabit ENet.
the interchangeability of hardware is important from an IT perspective. having parts laying around, or being able to repurpose a boxe's task is important most netoworks are still mixed usally containing some form of windows - the linux servers that get decomissioned can still be used for other tasks that require windows. the same is not true for sun machines.
In the end I'm not saying that linux categorically eliminates solaris from the world, that is not the case at all. I'm trying to make the point that Sun is playing in a world in which they do not have the ability to compete, one that linux and intel have licked. There are other markets that will have a need for some time for the high end n-way boxes that sun produces but not the areas in which linux is being adopted. Linux is not the one stop shop answer to all the licensing woes of the personal computing and IT world; but sun's "brand" no longer carries the same singular sense of reliability and depend that it once did - becuase there are alternatives. Change is slow in the IT world - with the dot com boom gone *boom* - saving money is cool, and budgets are slim. these all point to Sun >rapidly< rethinking their current direction.
In two to three years when the existing dot com then "legacy" systems come up for upgrade - intel and amd hardware will be very attractive from a price/perfomance standpoint AND they will have a *nix esq OS that is available for deploying all of their Linux ported applications. The choice, will be clear.
I've done as you asked:
/. a few months ago about the cost of ownership for windows, linux and solaris. solaris wasn't even in the running. on going costs were vastly higher than the other two platforms.
Sun v120 (There is no v150) 1U
SINGLE PROCESSOR CAPABLE @ UltraSPARC IIi 650 Mhz
1 Gig RAM
2x 36 gig SCSI 10K Drives
2x10/100 On Board-Nic
1x1 Gigabit Ethernet ( added to compete with dell svr.)
UltraSCSI II (No Hardware R
aid)
cd-rom
No redundant power supply
1 yr. wrty.
Price as configured: $5,757
DELL 1650 1U
DUAL Processor Capable w/ ONE processor @ P4 -1.24Mhz
1GB RAM
2 x 10K 36 GB Drives
Hardware RAID - UltraSCSI II On-Board
2x 1 Gig E-net On-Board
cd-rom
redundant supply available
3 YR. Warranty
Price as configured: $3,640
basically for every two sun servers - you could get 3 dell servers. they cost less to upgrade too.
under most circumstances i'd recommend that drives and supplies be redundant - it doesn't hurt and can only help. Sun charged 1K for the extra drive @ purchase time. Most notably sun lacked the one gigabit on-board ethernet.
A server's sole purpose is to push bytes and process. slower nic's = slower performance. The lack of mutli proc on sun's part isn't surprising - you have to pay for expandability. Also of note is the lack of redundancy i.e. Hardware RAID support and redundant power supplies.
The v120 is their "entry level server" that is clearly lacking in expandability, and reliability. As stated before - sun hardware at the web server level cannot compete. just as sun does a shoddy job with their software - they do the same with their hardware - and they won't even warranty their hardware beyond a year. Repairs aren't cheap.
As for maintainance - starting adim salary for linux is lower than that of sun. also, because linux is freely available - the likleyhood that a candidate has spent more time working with linux boxes in and out of a work environment is higher.
there was a study posted on
Sun has demonstrated a particular ineptitude in software development. For example Java has become a rather messy collection of API's. Some might call it "robust" however Sun constantly defines new API's touting them as "innovations". However there is no leadership in execution. Why is it that developers are forced to rewrite sun written code to enchnance performance and make their applications reliable? The truth is that Sun is very fond of "reference" implentations. I.E. they are half assed, untested and unreliable. To get really good perfomance other have to start from the ground up. Blackdown, IBM's 1.3 JVM comes to mind, Caucho's XSL Parser and there are others. Sun's leadership with repspect to software is really none at all. Setting up a portal and creating "standards" through API's isn't really adequite.
The restriction that I was refering to was primarily hardware. Intel hardware, the most common and definately most compelling hardware platform to deploy linux on does not scale particularly well. If a machine was needed that required 32 processors IBM or Sun would likely be the only two choices. It will be a long time before linux will be able to play in this arena. There will always be a need for these machines.