Slashdot Mirror
Sun introduces the "Sun Ray"
Doofuswrote to us about Sun's release of their newest effort to knock the PC off the corporate desktop.
The Sun Ray is essentially "a juiced-up monitor", and is a thin-client solution. Cost is 10$ per month for 5 years, or 30$ per month for a more powerful client. Not much technical details in the article, but we'll update with more links as these appear.Update: 09/08 01:15 by H :Thanks to Paul Tomblin for a huge PDF file with the tech specs.
It's really funny that after years of arguing the rising power of personal computers signals the death of mainframes, I'm now on the other side.
Since the early 1980's, the processor power of PCs has increased by a factor of almost 4000 and the strorage space by about the same amount. Where are the extra CPU cycles going? According to my NT task manager, 97% goes to the idle process! In a large organization, where does all the storage space go? Simple, hundreds of identical copies of the same applications such as Microsoft Office.
To understand why centralized computing resources like the SunRay have a chance, you must understand Total Cost of Ownership. In an enterprise environment, the cost of a network of computers is a combination of the price of the machines and software and the price of maintenence including factors such as software & hardware updates, periodic backups, and network administration. The upfront costs are dwarfed by the ongoing costs of support. The SunRay is directly targeted at reducing these costs.
My views of this topic have changed primarily because of the rapid bandwidth growth and improved stability of corporate networks. As 100kb/s and, in some cases, 1Gb/s connections proliferate, the differences between running an application locally and across the network diminish. But to the administrators, backing up 3 large machines is far easier than several hundred small ones.
Given one hour to live, the student replied: "I'd spend it with professor FP who can make an hour seem like a lifetime."
1.Nobody needs dumb terminals in today's
workplace environment. Real computers are
necessary, not slick looking terminals.
OK, why? That's an awfully general statement,
with absolutely nothing backing it up.
2.A five year commitment is too long a
technology commitment in today's marketplace.
Not when only barely functionality is on the
desktop. Do you constantly upgrade your monitor?
With thin clients, you upgrade the SERVER. So
long as the client has enough colors, high enough
resolution, and doesn't break, they'll last for
YEARS. The hundreds of PCs you upgrade every two
years for $1k apiece is that much money you sink
into the server. And ~$100000 will buy one hell
of a server.
3.This won't integrate very well with a
Windows-centric economy.
Well then, I guess we should all nuke our Unix
partitions and go to Windows then, if there's no
point in trying.
4.It doesn't just involve buying a thin client.
It also involves buying the server, the software,
the administrators to configure it all and the
technicians to train the masses
Umm... the cost savings from buying hundreds of
PCs buys you the server. You'd need to buy the
software ANYWAY, whether its hundreds of single
user licenses or one network group license. If
you don't hire an idiot, you only need ONE admin
for the server, and the whole POINT of these
devices is that you turn them on and go. A person
who needs training to use a monitor or telephone
shouldn't be allowed near either.
I think the problem is you're stilling thinking
small... just you, sitting at a PC. These devices
were made for groups of hundreds of people, a
level where one independent machine per person is
a nightmare... where you DO need dozens of techs
and administrators and constant upgrading.
--
Brandon Hume
hume -> BOFH.Halifax.NS.Ca, http://WWW.BOFH.Halifax.NS.Ca/
Brandon Hume
hume -> BOFH.Halifax.NS.Ca, http://WWW.BOFH.Halifax.NS.Ca/
Uh huh. I could tell you how many accountants here that write excel spreadsheets all day have PIIIs on their desk. They're not all that complicated spreadsheets either.
Let's be realistic here, who really needs a PC on their desk:
Now, which will be more expensive in the long run? You have a $10/mo/person thin client, plus the $10k for a beefy server. Or you have the endless upgrades of $2k/person/year plus the $10k for a beefy server plus the cost of moving machines around, fixing broken hardware, etc.
The artical implies that these devices are aimed at corporate customers, who will be attempting to reduce costs. This device is going to cost them $600 over five years which, although cheaper than a PC, is still a considerable investment. This does not have 50% of the functionality of a PC, yet costs more then half as much.
The other big problem I see is that coperates tend to like runnning their own custom software packages, and customising the standard ones. IS Sun going to allow these packages to be uploaded to its own execution servers (and provide the necessaryt security), or is it intending that corporates buy their own servers. If the latter is the case then the total cost is not going to be far short of a PC anyway.
I'm actually curious to see what the resource requirements are to support a bunch of these things deployed in a company. One box per 5 clients? 25 clients? 50 clients? Is the server end a web server, or is it custom software for Solaris, etc...
It will be interesting to see how Sun balances forcing people to buy servers (which they want to do) with integrating this technology with a business's existing servers.
Cheap PCs are cheap up front. The thing is, the hardware cost is nothing compared to the cost of maintaining the things. Sure, a well-managed *nix workstation is easier to manage than an NT box under SMS or Tivoli, which is easier to manage than a Mac, which is easier to manage than a Win95/98 PC.
/. testosterone-fueled lust for playing with computer guts isn't shared by most people, nor will it ever be. Nor should it. Most people want nothing more than a foolproof, zero-maintenance way to run a range of general-interest apps. Sooner or later, the PC as we know it is going to become something only developers and hackers will want. Everyone else will be perfectly happy to plug away on a ROM-based box with high-speed net connectivity.
But all of these things, with their varying hardware, their local filesystems, and in too many cases their local apps and OS, are a total money pit compared to running thin clients, whether they're pure terminals, or something with local CPU but no local disk-based apps and data, like this.
Past NC attempts have been underpowered, and viable apps outside vertical markets have been few. But at some point, large businesses will be more than ready for the right thin-client machines.
Besides, our own
Will Sun Ray succeed? Ehh. The odds are certainly against it. Will something like it succeed in the next couple of years? Yup.
Sun Help's Rumor Page contains some corona references.
I feel this needs to be reiterated in the main thread.
I'm reading so much about how much easier terminals/network computers/etc are to manage. And I'm not disagreeing in the least. The time and costs required to manage a bunch of PCs grows exponetially with the number of boxes on the network. It's a given.
My problem is that everyone seems to be looking at this with a very narrow point of view. Which is also to be expected. The readers here are primarily technically proficient, intelligient people. And I'm sure that quite a few of them are full time sysadmins as well. And from their perspective, easier management and reduced cost of the network is top priority, as it should be, for the most part.
But you need to consider the welfare of the entire company (and this message is directed at business computing, since I'm sure we can all agree that none would want to give up their home PC for just a terminal).
If you have a company full of glorified data entry personell who don't have the ability/desire/intelligence/etc to use more than one or two applications, this works well.
But I believe that the added costs of hiring more adept people, giving them the tools and resources (PCs) to use their abilities will give the company as a whole a competitive edge. THIS is the correct reason for moving from a mainframe/terminal setup.
I'm sorry, lowering IS cost is meaningless if it means reducing the employees ability to do their job effectively. This of course requires the best employees possible, which is another topic altogether.
I'd much rather have one or two intelligent people with atmospheric salaries who can creatively solve problems than 30 entry level people who require a lot of hand holding and attack problems brute force. And if it costs more to maintain the network and to give them PCs for them to do what they need to do, so be it. It's a drop in the bucket compared to what they make and what they make the company.
As I read this article, this device is basically an X-Terminal. Which just proves that X-Terminals are great devices that aren't used nearly often enough.
:)
I used to run a 5000+ user UNIX environment, with over 500 X-terminals. Alone. By myself. With time to spare. The durn things never broke
The downside to X-terminals has been that they tend to have an up-front cost almost as high as a workstation. It could be that Sun Ray will fix this.
-- Slashdot sucks.
I mentioned this a few stories back, in the StarOffice discussion. Sun is not flexible enough to compete in the PC market, so what do they do? Try and manipulate the market to fit their business model. It's the network computer concept that they keep trying and failing with.
Sun needs to realize that people like their PCs. Whether you run Linux, Windows, MacOS, BEOS, whatever. We moved away from the mainframe/terminal paradigm for a reason.
This drives Sun crazy, since it threatens their extremely high margin server business (talk about price bloat). Where do you think Sun gets all these millions to buy StarOffice and give it away free? Or put so much development money behind Java?
Sun is robbing people for their servers. And they'll continue to do it as long as they can.
What many of the thin client vendors (and many in the Open Source community) miss out on is that not every business is using computers to run MS Word, Excel, and IE. We're running a scheduling a business package on an Alpha Microsystems box (?) and using dumb terminals. Sadly, I can't just use VT100 emulation, as the emulation mode is called AM-65. Looked high and low a few months back, and the only terminal emulator I found is made by the SOB's who make the system. Yes, we are looking to replace it, but the funds to transfer the information from the old system to the new just aren't there.
Then there is the vendor of our computer based medical records system. Unhelpful. Totally MS based. No chance of Open Source (we are a "development partner" and we can't even get the source. Not that there are any programmers here, but it's the thought that counts. What we do is develop templates that are then passed around to the other users without credit being given). No chance of a Linux, X (in general), Wince, Palm, or MacOS port.
So what does that have to do with these new terminals, or any thin terminal? Quite frankly, I'd love to use them here at the office. Doctors are not the most technically savvy folks. Sure, they can use the latest laser to burn away part of your colon, but I have yet to meet one who could program their VCR (lest the MD's flame me, I've been around docs since I was born. Unless you're about 60 years old or so, I've been around more docs than you) Anyway, thin clients would be a lot easier to manage, and would give me more time to start my business from my cubicle. But the numbers don't make a damned bit of sense. For just a tiny bit more than $10/mo, I could lease a MUCH better machine (even if it's saddled with NT, which, once running, is much better than 95/98). Of course, I'll be leasing for only three years, as a five year lease for computer equipment is foolish. We've got some stuff due to be finished with the lease in about six months, and the leasing companies are hard pressed to give us a buyout, as there isn't much of a market value for 486/DX4's and Pentium 66's.
So while thin clients are nice, the lack of supported applications is sad, the price is absurd, and it just doesn't work. Thin clients work quite well with CLI's, but until someone has a sanely priced graphical client, what's the point? Wyse and Sun have missed the boat. If they are going to make this work, they are going to have to work with vendors and developers to come up with more web enabled apps, java apps, and other tools that are not as mundane as word processors.
Jesus was all right but his disciples were thick and ordinary. -John Lennon
They are _extremely_ dumb, not even X terminals. Instead you have a terminalserver, that runs one X server for each SunRay terminal. Then the bitmapped graphics is transfered over the network, in some compressed format, all the terminal does is send the keyboard and mouse events the other way, and put the graphics in the framebuffer. Exactly like VNC and Citrix, not something that sounds very intelligent.
I have only used them briefly, but they actually seem very fast. Ofcourse I don't know how they stack up under heavy load. Don't expect fullscreen MPEG on them though
Tech details: 1280x1024 @ 76 Hz
24-bit colors
10/100 Mbit Ethernet connection
Composite video input
Stereo audio out/Mono microphone in
4 USB port
ISSO approved smart card reader
The setup is 100 terminals, with 50 each on a Sun250 Terminalserver (Dual USparcII, 2G ram)
These only do the graphics, 50 X servers on each there is a HPC6500 for the CPU power with a couple of E10K to come.
I don't know if the page describing the new setup is available from the outside but try:
Databar update
Morten Olsen (not AC)