Led by upstarts such as RLX Technologies, some of the large system vendors have jumped on this innovation within the server market. Sun’s Subodh Bapat, CTO for Sun’s volume systems products division, recently sat down with the IDG News Service to discuss his company’s recent Blade server announcements, and its vision for the next generation of the technology.
Sun Preaches Blade Servers
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Contrary Computing?
Contrarian Circuits?
With almost every aspect of Sun’s products, they are doing everything the opposite of the rest of the industry.
It is difficult to think that all computing needs will fit into a blade form factor. And the timing of Sun’s move to ‘form factor over functionality’ seems far too early in the server market.
In the server market, the software that runs on the hardware is the main weakness. There is no such thing as plug and play. Form factor is not an issue compared to getting everything to work. IBM knows this. Sun … is wandering around in an eclipse…
I’ve said it once and I’ll say it again, SUN need to bundled Solaris and iPlanet all into one box and sell a unlimited client license for $599-$699 and include 6months of free technical support via the telephone/email. No one in the right mind is going to pay several hundred for the operating system, several thousand for the webserver, then several thousand for the directory server then several thousand for the mail server etc etc. Take a leaf out of Microsoft’s book and give the customer a one stop solution in a box with a nice price tag. That is what people want, not a lego set where by the customer has to hunt around and hopefully find the right combination of software to work together.
Why, CCCC, that’s a great idea! Sun can make it easy for the customer! Wow, that’s executive level stuff. Make that “C-level”. Amazing.
And your unbridled optimism regarding Sun’s intelligence almost brings back a memory or two of the good old days before the dotcom talent dilution.
However, in the real world, Sun doesn’t seem to understand listening to their customers, respecting their customers, or building products that actually solve problems for their customers. More importantly, it is hard to tell if Sun even wants to do these things. More and more it seems Sun wants to invent some crazy technology and force feed it to the world, not realizing Sun has very little mojo left in the modern age of computing.
When it comes to software, all you hear from Sun is “Java is the be-all end-all solution to all your computing problems”. Sun doesn’t understand that any one computing language has a finite lifespan and that positioning your company as a “one language” company creates a vast polarizing effect. And Sun doesn’t even know how to take advantage of that effect. Sure, Sun has sold lots of hardware in the past to make their Java bloatware run at acceptable speeds, but Sun has made almost no money on Java software.
Back to the topic. Yes, if Sun were to bundle together everything it takes to build Sun software solutions into one affordable package, it would be great for developers. Now just don’t hold your breath waiting for it to happen…
Back to the topic. Yes, if Sun were to bundle together everything it takes to build Sun software solutions into one affordable package, it would be great for developers. Now just don’t hold your breath waiting for it to happen…
I sent this idea to a number of SUN execs. The sooner SUN dies, the better, then maybe we’ll see IBM’s UNIX business give Microsoft a run for their money.
As for SUN, they’re the maker of their own destiny and if they think that blaming Microsoft for their tail of wowes is going to help them, well, they’re wrong.
When I was studying information technology, I always wanted to work for SUN, now unless I can enter as a manager, grab SUN by the nackers and give the organisation a bloody good shake up, it isn’t worth joining at all.
I sent this idea to a number of SUN execs. The sooner SUN dies, the better, then maybe we’ll see IBM’s UNIX business give Microsoft a run for their money.
Sun is already planning on doing this with Orion…
I don’t know how you sent your “idea” to “a number of SUN execs”, but that sounds absolutely ridiculous…
Sun is already planning on doing this with Orion…
I don’t know how you sent your “idea” to “a number of SUN execs”, but that sounds absolutely ridiculous…
Why does it sound ridiculous? because it is the bloody obvious? who the heck is going to go to StarOffice 6.1 as their Office solution? are they really going to convince people that they should give up a whole heap of functionality and retrain everyone with some office suite developed by a company in the financial crapper?
SUN had their chance, they’ve blown it big time. Sooner or later “loyal” SUN cronies will see the writing on the wall and evacuate. Atleast with IBM’s websphere solution there is some vision in where they’re going to and actually focused on delivering value to the customer.
i’ve just had a look at their offering, and I want to wait and see what they have to offer. I am basing my conclusions on their past behaviour, however, if SUN is back on track and this does turn out to be a winfall, it will be an interesting situation.
Bascule, I know I sounded a little harsh, however, I personally don’t want to see a company like SUN go down the toilet. They have great technology, great engineers and a good fighting spirit. I’m just a little concerned when I hear things such as “Linux PC”, and other non-core things being developed.
SUN needs to focus on what it is good at, namely, producing top quality, reliable, scalable workstations and servers. They should try to leverage their SUN Blade 150 as an alternative to the desktop PC found in organisations. I’ve just had a look and to be complete honest, there is only a very small difference between a IBM corporate PC/workstation and a SUN Blade 150.
I would love to see SUN push the SUN Blade 150 + Solaris + GNOME 2.x + StarOffice + Evolution 1.4 as a viable alternative to the PC in the corporate world. Yes, they would need to do some cost cutting in production such as out sourcing, and bring UMC and TSMC into the chip production side (which should slice a few 100 off their price), and push that as part of an over all solution.
I’ve worked for a Sun VAR in the not-so-distant past.
Solaris costs $100 (at least to our customers).
It’s usually free with hardware.
What I was getting at is this. Sure OS is cheap, however, why should one feel that they should then shell out more money for things that are included with other operating systems for free? After having a look at Orion, maybe this will be a good reply to Linux and Windows in terms of value for money.
As for the cost of Solaris with hardware, it is pre-installed onto the machine, however, you have to purchase the media seperately. The media isn’t included with the hardware.
I am not too much into blade servers myself – a good SMP solution is far easier to manage and software runs more efficiently on it – but of all blade-kind offerings, Sun offers the better integrated and most easy to administer (as usually with Sun servers, with their richness of RCS, LOM and other advanced administration solutions).
If you actually read the article and have been keeping up with Sun’s product announcements.. you’d understand their direction.
Sun has reduced the pricing of it’s equipment and increased the speed considerably. The 1.2GHz CPU costs the same as the 900Mhz CPU and is 60% faster than the 900Mhz. Sun will have the UltraSparc IV out this year which will double CPU power without needing to upgrade to newer systems. That is good protection of a company’s investment. Company’s don’t want to keep buying new PC crap every other year and spend most of the time in-between implementing it.. only to have to replace it with whatever is the latest. Sun has always listened to customers in this department. Sun makes good quality equipment. That is why they sold over $16 billion in servers last year and the used Sun market sold over $9 billion.
Sun is not going to leave the UltraSparc or Solaris product lines, they are the key differentiators for Sun. Sun makes x86 boxes at the LOW end, and most people use them as appliances.. Firewalls, SMTP relays, cheap web servers, etc. And the blade systems are excellent for horizontally scalable services, like Apache. On the other end of the spectrum, Sun will continue to make it’s big iron servers like the 15K. The point being that Sun has products to fit customer needs from the low end to the very high end. The advantage is that Sun has a stack of software that runs across it’s UltraSparc product range. All of that is built upon Solaris. Solaris has many features built-in that you won’t find in Linux, like (RBAC, DR, SLVM, etc.). With zoning in Solaris 10, Sun will be able to increase server utilization and have more mainframe like abilities built-in. With project Orion, Sun will include it’s Sun One product line with Solaris and offer very competitive pricing for those components as you use them beyond the included licensing.
Sun as come a long way and they are still competitive.
And for those who don’t believe me, I’ll give you a real example. I can replace my old Sun servers with new V480’s, V210’s, and V240’s an only pay a few $100 hundred dollars more a month on my lease. The company I work for is replacing it’s old Sun hardware (E450’s) for Oracle with a 6800, V480’s, and V120’s in a lease that costs less than half what we pay for our Mainframe every month. Once we replace our Mainframe apps with new Oracle apps, there won’t be a need for the Mainframe.. that’s a lot of money to be saved. The key thing is that you have to go through a good Sun VAR and work with Sun to get the pricing where it needs to be. Any major purchase a company is going to make, takes good negotiating, and Sun is a great company to do business with. Don’t take the pricing on Sun’s webpage as the lowest price for the configuration you want.. because it isn’t.
And when it comes to things like Oracle, you need systems that can handle I/O, large amounts of memory, and are fast. All of which you get with Sun hardware. While x86 cpu’s may be faster in spec benchmarks, they can’t hold as much memory or push as much I/O. And that means you’d need to buy more x86 gear than Sun gear.. which means greater licensing costs.
Now Oracle is all in favor of people switching to Linux and buy tons of x86 servers, because it means more licensing money for them:) So the savings aren’t there in the long run with x86 gear. With x86 gear, you end up spending more money because you have to upgrade to newer systems soo quickly. And when you add on M$ software, it can get very expensive. With Sun, you buy hardware that can easily last 5 years with only upgrades to CPU’s, Memory, and storage. That’s a better deal than buying new x86 systems every other year.
Sun is turning itself around by focusing on it’s hardware and software products. Now it just needs to get the message out to it’s customers. The biggest hurtle for Sun is that it’s always been horrible at getting the message out to it’s customers. That is why you’ll find soo many companies running old Ultra 2’s and Solaris 2.6.. they haven’t upgraded because they don’t know the benefits and most of the time don’t even know what’s availible. Sun has always built high quality systems that last. It’s a strong point and it’s a weak point for Sun. Most customers who have been running Sun gear for a while don’t upgrade to new systems often, because the old stuff they bought from Sun years ago is still working. In some companies, people don’t even know that key parts of their business run on old Sun gear sitting in their datacenters.. it just works. And that is what Sun really has to face.. there is a huge market to refresh with new gear, but that market isn’t hearing the message. Sun needs to push on the advertising and hold more conferences to educate the market.