Hi,

What do you feel is/are the likely future infrustructure(s) for the hosting industry?

Still the single or dual x86 based CPU rack mount server that's so prominant now?

A blade server? A mainframe running *nix partitioned into thousands of virtual servers? What?

Please elaborate anyway you want -- be it a timetable for the changes you see, operating systems, control panels, customer service, what it will take to be a successful hosting company -- whatever makes up the future of website or web service(s) hosting as you think it's likely to be.

Thanks for sharing honestly -- and please, no April Fools jokes on this thread; thanks,

Louis

Hi,

Does no one have any opinions or experiences to share on this issue? Wondering since 39 views but no replies.


Anyway, take care one and all,

Louis

Only a few companies will ever even consider using mainframes or blade servers just for the fact that they cost so much. If IBM or other large players such as Microsoft of Compaq etc were to step into the market, then yes, they'd use something like that, but for 90% of the industry, that type of equipment is not feisable.

Personally I'd love to get a massive Sun Microsystems mainframe, but the cpu's alone in those puppies are several thousand dollars, and i'd be broke by the time I got the first customer on it, the amount of time it'd take to get a server like that to capacity and to get it to where the server was actually making money instead of loosing money would take a very large company and a big advertising campaign.

Someone who pulled 20-30,000 virtual accounts a month might could justify using a large expandable mainframe but thats about it..

A cabinet clustering system so you can start off with just 1 1u server and just pop another one in the rack hotswap when you need more power.

Blades. Most definately, in a highly available clustered environment with NAS or SAN storage.

Not for the $10 a month hosting account, but for serious, scalable, resource on demand web hosting.

-t

I think all them makes sense.

If they standardize a format for blades and integrate standard desktop hardware I can really see the market taking off. What are current blade sellers doing 8 or so systems in a 4U chassis? That's pretty damn good rack density. I can see the benefits in both hosting applications and the corporate desktop enviroment.

Mainframes could definately bring value to the hosting market but I don't see a lot of the little guys bieng able to afford this. Maybe we'll start seeing larger companies purchasing bigass sunfire boxes and selling off a percentage or two of the resources to resellers. The scary thing to me is that the smallest amount of downtime would cause SERIOUS issues for thier customers. I understand the redundancies built into mainframe type systems but still, can you imagine if one screwup takes down 500,000 customers?

Hi,

I really wonder about the economics of the new server options such as blades and virtual servers on a Linux mainframe in the near future (1-3 years out). I'm wondering what this will do the the cost structures of the dedicated and managed server markets -- and what this will mean for and how it will affect the hosting industry.

After everything's said and done, I wonder how the cost structures will compare between the single or dual CPU rack mount units of today and the blades and virtual servers running on Linux mainframes of tommorrow. What think ye?

Yes, mainframes do cost a lot. However, divide this cost between tens of thousands of Linux virtual servers -- in effect tens of thousands of load balanced virtual servers on a mainframe -- and I'm wondering if this will result in some heavily capitalized companies moving into the hosting market space offering hosting at the $10 a month and less level.

Uptime with mainframes is top notch in a properly run data center. In 13 years of working as a mainframe programmer/analyst on large, mission critical systems running on mainframes run by very large companies, I can count the number of times the mainframe has been down on one hand; those boxes just crank.

Even if there are cost advantages to blade farms and Linux mainframe virtual server farms, I'm also thinking that there will still be a market for smaller hosting companies selling space and bandwidth on more conventional single or dual CPU rack mount servers.

It seems to me as long as you can sell the space and bandwidth a rack mount server provides -- to either retail customers or resellers -- for more than it cost to own, maintain and administer the server you have in a co-lo -- there will continue to be a market for the smaller more conventional hosting business -- even in the face of an onslaught by some "biggies" selling tens of thousands of cheap virtual domains on a Linux mainframe.


What think ye?


Thanks for sharing everybody; I think this is an issue that will affect this industry in the future.

Louis

Well, speaking of the future, who's ready to switch over to IPV6?

Who's studying VR technologies? :)

Can't see us using blades. Each server needs to be robust, protected against disk loss and a performance vehicle.

Blades are all about mass hosting in small space. Not our market.