April 15, 2009 at 3:37 pm
Comments posted to this topic are about the item Dreaming of Clouds
April 16, 2009 at 4:58 am
Please forgive me for being naive, but isn't much of what was discussed already in place? From that I mean, our users have no idea where the physical machine is, or what it is, nor do they care. All they know is there is an ODBC connection called "IAR_Warehouse". And then even that is sort of foggy because of DNS aliases.
The argument that developers won't care where the services are isn't much better, for they all use the aliases as well.
I guess it all boils down to the fact that at SOME point, you still have to have people that deal with the machines themselves and not some abstraction. The only thing that I see the cloud initiatives doing is pushing that responsibility to another group of people.
(I would like to soften the coming criticisms. I'm working on very little sleep today and the ankle I broke 5 weeks ago is particularly sore from "working" on the fence.)
Honor Super Omnia-
Jason Miller
April 16, 2009 at 6:49 am
I agree that a lot of this either exists today or will int he very near future. Seems to me much of the same ideas are being floated by the VMWare folks and their Virtual Data Center OS ideas. The ideas of provisioning more resources automagically when they are needed (vMotion?). Having no idea what physical machine is actually running your database, VMWare Clustering etc. etc... Granted this is all for your internal intranet type of thinking, not so much for including services provided from the web.
EDIT: yes I know MS is talking about the same types of things, I just am more familiar with the VMWare product offerings... please don't shoot me for only talking about VMWare. I don't work for them I just use their products.
April 16, 2009 at 7:38 am
It sure sounds interesting. And a lot of it is definitely either already here or on the horizon and visible.
As always, it'll mean even more specialization in IT. How many companies have "virtual system admins" now? Some title like that will become standard as these things become more common. Right now, the function exists, it just doesn't have a title, and usually isn't "single hatted", but is a shared duty for the server admins.
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
April 16, 2009 at 7:49 am
From the point of view of someone who's seen a lot of hype about cloud computing but knows little of the substance...
I'm very cynical about the whole thing. I can't imagine it making our lives much easier (maybe I need a better imagination 🙂 ) but I can think of plenty of ways it could make our lives more difficult.
Got a performance problem? Good luck finding out what's causing it, or even replicating it consistently in the cloud. Want to improve performance? You can start upgrading all your machines or adding new machines in the hope they'll make a difference, but I'd much rather have a definite hardware configuration to benchmark against & improve upon.
It seems to be the cloud will add complexity & remove a great deal of control. I probably need to do my research, possibly I just need to look at it again with the aid of more chocolate. In the meantime I'm happy to be put straight 🙂
Ian
April 16, 2009 at 7:56 am
I think the whole idea is that there will be a set of people/systems that will monitor hardware for you and you won't have to worry about it.
If you're having a hardware related performance issue, that should, theoretically, show up in a trace, alert, whatever, for the person/people/robot, saying that such-and-such CPU is overloaded, or that the hard drives in SAN 123 are running a bit hot, or whatever. Then they handle the specific issue, without you having to know about it, much less do anything about it.
In other words, it's just yet another layer between you and the hardware. Same idea as the OS and drivers and such, plus the SQL engine, making it so you don't have to write your own interface to the hard drives or the screen, you just tell it to insert data X into table Y. Just adds one more layer to the whole thing.
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
April 16, 2009 at 7:58 am
icocks (4/16/2009)
Got a performance problem? Good luck finding out what's causing it, or even replicating it consistently in the cloud. Want to improve performance? You can start upgrading all your machines or adding new machines in the hope they'll make a difference, but I'd much rather have a definite hardware configuration to benchmark against & improve upon.
Great point(s)!
The bottleneck will be difficult to diagnose. So the IT guy (whatever they'll be called) tries to replicate it, doesn't see it on his run, just dismisses it as end-user fiasco. Then end-user doesn't get satisfaction, gets p!ssed off, and in turn blames the lousy support guy, and the merry circus continues.
Honor Super Omnia-
Jason Miller
April 16, 2009 at 8:00 am
Some of this is here, some isn't. Now users don't necessarily need to know the physical server, but in some sense they do. They put in a server name, which is somewhat tightly tied to the physical server. You can go with DNS, and that helps, but it doesn't help with the motion.
VMotion works well for many things, but not databases. The few people I know that successfully put a database on a VM don't have it floating.
The idea isn't new, but it's visible as something that could be achieved. I'm not sure a lot of what is done in technology is really "new", it's re purposing old ideas, using them in new ways, and that can make things much more productive.
Someone does need to know how things work, but that should be less and less people, which allows for more advancement in society. Think about how many things we don't need to know how to do, or how they work, that our parents or grandparents did.
April 16, 2009 at 1:43 pm
[font="Verdana"]Sorry Steve. I'm going to reserve judgement. Most of it (to me) still seems to be pie in the sky dreaming. Yep, great to have something to aim for. In the mean time, it doesn't change anything for me. I can't architect with hypothetical pieces. :)[/font]
April 16, 2009 at 3:16 pm
Jason Miller (4/16/2009)
I guess it all boils down to the fact that at SOME point, you still have to have people that deal with the machines themselves and not some abstraction. The only thing that I see the cloud initiatives doing is pushing that responsibility to another group of people.
Agreed. The "cloud" already exists for most people, perhaps not for most DBAs - at least not yet. To me this is more like a further refinement or division of labor within the DBA ranks compared to what exists now.
Much like users and developers today never see the servers or databases they use or care where they are or what kind they are (as long as they work) - some DBAs will become more removed from all that also. And some will work in "up in the cloud" where they actually can see and touch the servers and work more directly with the databases.
April 16, 2009 at 5:00 pm
Often we are mistaken about how a new technology will actually be deployed and utilized, and the cloud is probably no different. Many comments I've seen about this paradigm assume that everything else will be as it is today, except the database is somewhere in the cloud. But it's unlikely that a typical web application with it's supporting DB would ever do this. In most instances, either the whole application is in the cloud, or none of it.
Are there good reasons for putting a database in the cloud but without it's associated application? Yes, it could make sense for things like large scale data migrations between companies/vendors, replication and archiving, environments where various distributed applications use the same data, etc.
LinkedIn - http://www.linkedin.com/in/carlosbossy
Blog - http://www.carlosbossy.com
Follow me - @carlosbossy
April 16, 2009 at 5:02 pm
icocks (4/16/2009)
From the point of view of someone who's seen a lot of hype about cloud computing but knows little of the substance...I'm very cynical about the whole thing. I can't imagine it making our lives much easier (maybe I need a better imagination 🙂 ) but I can think of plenty of ways it could make our lives more difficult.
I'm right there with ya. Though, my issue is less with the technical side and more on the business side. For the majority of us who will be stuck using the "Big" cloud (i.e. the nebulous SQL Services on the wire) we will be at the mercy of the 800lb gorilla. Think ATT circa 1984 or Microsoft circa now. They can do whatever they want, suck as much as they want, and charge us whatever they want. Got a performance problem? Pay us more and it will go away. Data breach? Sue us. We have an army of lawyers on retainer and an iron veil of secrecy to protect us. Can't connect? Your network must suck. Oh wait, our network sucks. Eh - still your problem.
Cloud providers are being framed as new utilities. As the Cloud concept progresses I expect to see providers go through the same maturation process. There's lots of competition at first, prices are great and service is wonderful. Then comes market saturation. Followed by consolidation. Then there is a monopoly (or cozy duopoly at best). Prices go up, service goes down, we all get shafted and some dickwad - who isn't you - ends up a billionaire. Then the gov't steps in a regulates the whole thing into oblivion. We all get sick of the Cloud and move on to the next big thing. Lather. Rinse. Repeat.
I guess that's why they call it progress.
James Stover, McDBA
April 16, 2009 at 7:10 pm
Sorry Steve. I'm going to reserve judgement. Most of it (to me) still seems to be pie in the sky dreaming. Yep, great to have something to aim for. In the mean time, it doesn't change anything for me. I can't architect with hypothetical pieces. 🙂
This is one of the sponsors of Seattle Cloud Camp there is a page generating API and automated deployment tool and media hosting API. All of the above is simple technology of limited use to very few companies if you are not in the media business. So Cloud is clouded by limited features, maybe the Cloud people needs to talk to Microsoft because we don't accept what Microsoft provide if we cannot test drive it, break fix it and may be run the Reflector on it.
😉
:Whistling:
Kind regards,
Gift Peddie
April 16, 2009 at 9:05 pm
So, this article is about clouding a blue sky! Or are we trying to put some blue sky on the cloud front?
"Blue Sky" refers to looking at things outside the box of existing situations and environment, or even known technology. It is kind of a, "If I can have what I want and need, what would it look like."
The "cloud" is about abstracting the software from the hardware (very generalistic statement here).
So, it still begs the question: What does clouding it up have to offer? Maybe a little rain will help things grow?
It doesn't make things simpler or cheaper. It may actually make things far more complex. But it also may open up some interesting opportunities that were not there before.
😀 😛
To me, the cloud still looks quite wispy and projected (or protracted) as the wrong tool for the bulk of private and governmental IT needs. I am still looking for better, so I will focus on clear, blue skies until something comes along that really shakes the house. 😎
Viewing 14 posts - 1 through 13 (of 13 total)
You must be logged in to reply to this topic. Login to reply