April 17, 2009 at 8:18 am
Currently my experience has been with small companies and SQL Server has been used as a backend to specific applications where the app is being pushed out to the cloud. (cloud meaning internet based so people could use it on the road.) The need to expose the DB to the cloud, in the cloud, from the cloud is simply not there. What more is needed from the app then a good connection string in this type of setup.
I do see a neat idea when we will have one computer environment where we simple add and subtract hardware from a mass of servers and they all act together. We simply load the applications, from SQL to SAP to word and excel and the computer environment manages all processors and memory and storage. This is where cloud computing would really take off. In this version SQL does not need to change fundamentally to succeed.
John Burris
MCITP Database Administrator
[font="Tahoma"]John Burris:hehe:
MCITP Database Administrator[/font]
April 17, 2009 at 9:37 am
I'm with John on this one. The best I can gather (limited that it is), this cloud thing is a unified application/server/service environment, whether Internet or intranet. Whether the OS is installed on one machine or 100, with one CPU or 64, one hard drive or 500TB RAID won't matter. It's the true hot swappable system. One whole box could go down, and the OS just re-allocates the resources until a new box is added. It seems this whole cloud idea is around 24x7 availability, and the Internet angle is just about where (when and how already covered). SQL Server would just be another app on this system, just like any other. This push to get SQL server on the cloud seems to be trying to skip the OS management part, which seems backwards to me. Get the OS to do the cloud part, and the rest is just connectivity (security included).
I like the idea as a whole, and look forward to seeing how it evolves.
Dan
April 17, 2009 at 10:02 am
GSquared (4/17/2009)
... Given the chance at a free version, sort of Cloudy SQL Express, I'd at least give it a shot.
I agree here. I need a hands on to really get a feel for the technology. I'd like to see how it all fits together without having to find an employer ready to invest. Even if it was sort of single instance, single cloud-puff limited architecture, I'd love to see something like this.
😎 Kate The Great :w00t:
If you don't have time to do it right the first time, where will you find time to do it again?
April 17, 2009 at 10:37 am
katedgrt (4/17/2009)
GSquared (4/17/2009)
... Given the chance at a free version, sort of Cloudy SQL Express, I'd at least give it a shot.I agree here. I need a hands on to really get a feel for the technology. I'd like to see how it all fits together without having to find an employer ready to invest. Even if it was sort of single instance, single cloud-puff limited architecture, I'd love to see something like this.
This is available now in pre-retail stage (not sure if it is alpha or beta). It is a very limited capability set that focuses more on services and very light on SQL Server.
April 17, 2009 at 10:42 am
I don't think it is viable inside of the next 10 years. Even if we are talking about only internally housed clouds, as this topic lightens the topic with, it is a major reconstruction of application intercommunication and development.
Being able to wrap in 3rd party software that would also have to upgrade is a huge issue.
LOB DBMS's that are not already suited for internet business will not be going there anytime soon.
The real issue is at the Exec level, where the sales pitch sends IT into continuous cycles of "revisit the topic because it sounds cool".
April 17, 2009 at 1:16 pm
You know, like would Microsoft have a global version of TFS will ALL of their source sitting out there in the Cloud?
Not on planet earth but TFS running in every company is the plan one pesky problem all the nice TFS features comes in SQL Server Enterprise while you buy TFS with SQL Server Standard.
Do you think Steve Ballmer keeps a copy of Microsoft's business strategy on Skydrive??" Self is doubtful. At least not in the public Cloud - which is where 99% of us will be stuck in the real world. Like anyone without a billion bucks will have their own Cloud.
These are some of the problems that will keep sane mid size companies out of the Cloud.
I will happliy use Google Maps and Skydrive and Facebook and all the other inane Cloud applications. But I'm thinking until the 800lb industry gorillas start eating (and gagging) on their own Cloud "dog food", I will be perfectly content to keep all the really important stuff in my protected silo, far away from the Cloud.
That will not happen soon because the current features are so limited most is just grand CMS hosting with veneer. You need dog food to gag on dog food there is none in the interim.
😉
:Whistling:
Kind regards,
Gift Peddie
April 17, 2009 at 3:02 pm
skjoldtc (4/17/2009)
Whether one likes the idea of cloud computing or not, it is coming and we need to be prepared. It will not be a DBA's decision to use a cloud-based database it will come from an executive who went to a high-level conference discussing the potential benefits.We risk becoming obsolete in the marketplace if we ignore or dismiss cloud computing. Steve, I appreciate your editorials this week and it had prompted me to begin learning as much as I can about "The Cloud".
That and my boss told us that the previously-mentioned executive wants us to look at it seriously. Thanks again for the timely writings.
This about sums it up for me.
It's coming - actually it's already here. It's not for every app or data set, but does and will work fine for many purposes. It's not an "all or none" proposition. Bias for or against "the cloud" will not help advance either you or your company's prospects. DBAs would be best served to not dismiss it, but rather get ahead of it, and be the one in your company best positioned to help your company make the most of it, when applicable, while mitigating the risks.
April 17, 2009 at 3:37 pm
It's definitely a topic of discussion, and worth debating and understanding. Virtualization has come quickly, and it has failed in a number of places, because it wasn't well thought out. I know there were people that were excited about it and became soured because of poor performance.
There were also people that had it pushed on them because they had no specific reasons or citations why it shouldn't be done.
Whether you want to embrace it or not, you ought to educate yourself and have specific and documented reasons why you want to do it or don't want it anywhere near your SQL Server. At some point, someone will ask you about it, especially if it gets built into SQL Server.
April 17, 2009 at 4:47 pm
Bob Abernethy (4/17/2009)
It's coming - actually it's already here. It's not for every app or data set, but does and will work fine for many purposes. It's not an "all or none" proposition. Bias for or against "the cloud" will not help advance either you or your company's prospects. DBAs would be best served to not dismiss it, but rather get ahead of it, and be the one in your company best positioned to help your company make the most of it, when applicable, while mitigating the risks.
The "cloud" has been here since the internet came to be, and it has grown and continues to grow in capability. I also agree it is not an "all or none" proposition. I am neither for nor against the "cloud" itself. What I am against is the improper level and amount of hyperbole that is stirring on the subject. It is suited for some things, it is not suited for many things. The cloud has virtually progressed one or two steps, but the hyperbole makes it sound like it is 100 or 1000 steps further. Many tech types don't even know fully what or how it impacts, and this is only one of many concepts and training that DBA's need to keep up with. Honestly, in most cases, efforts spent in cleaning, consolidating, and optimizing current internal systems will produce far more ROI than a move to the cloud will. The hyperbole is trying to sell a market that does not yet exist except in a specialized and narrow segment.
It is like saying, "If you sell it, they will come, and then we can build it, fix it, and charge them even more."
The cloud is not suited for core LOB needs, unless those needs are based directly in internet facing needs. Internal VM, cloud, and any other form of centralized computing also has limitations, that must be weighed carefully. "[Being] the one in your company best positioned to help your company make the most of it..." applies to many facets of being a DBA, many of which are far more pertinent than the hyperbolic cloud pitch. The cloud is useful and adviseable in some instances, but not nearly a full sweep. As a metaphor, it is still far closer kin to Web Hosting than it is to Data Warehousing.
"Sounds cool" does not equate to "good business sense".
April 17, 2009 at 4:56 pm
Steve Jones - Editor (4/17/2009)
It's definitely a topic of discussion, and worth debating and understanding. Virtualization has come quickly, and it has failed in a number of places, because it wasn't well thought out. I know there were people that were excited about it and became soured because of poor performance.There were also people that had it pushed on them because they had no specific reasons or citations why it shouldn't be done.
Whether you want to embrace it or not, you ought to educate yourself and have specific and documented reasons why you want to do it or don't want it anywhere near your SQL Server. At some point, someone will ask you about it, especially if it gets built into SQL Server.
This is a reasonable approach. Know what it is and where it tastes best in the IT pie. And be ready to pacify with examples and documentation that starry-eyed arm-flailing exec who just got zinged by the latest sales pitch.
Either that or just go with it, ride the wave full tilt, and jump off right before it comes crashing down... ack... I am getting Dot Com deja vu here...
April 17, 2009 at 6:54 pm
Interestingly it appears that we're at another round of abstraction with things moving out of the data center. The "cloud" seems to be such an imprecise term depend on for business critical data. Kind of like a box with a "magic happens here" label defining your system architecture.
With that said, I do see the benefit of abstracting the data layer -- I just don't believe in magic boxes or magic clouds because when my boss asks me why things aren't performing I'd like to be able to respond as to whether or not it's a database issue.
The next few years will be interesting...
April 18, 2009 at 9:12 am
There's no magic, and I hate to see that. But there is a level of abstraction that can make things more manageable. That's what computers have done.
The "cloud" definitely needs more fleshing out, and that's one think I was hoping to do. Get more IT people to think about it, better define it, and not let management run off with some "hype"
May 8, 2009 at 9:26 pm
The US Government have defined what is Cloud computing because we are making fun of the Vapor ware we can only make fun of SAAS(software as a service) so the Cloud rain makers can continue because two sections are now no engineering required.
So I am hoping for better definition of what is PAAS(platform as a service) because when last I checked Java, Python and .NET are software development languages. And the web hosting companies are now in a separate category which is no new engineering required. Happy coding BC(before cloud)
http://cloudcomputing.sys-con.com/node/954002/mobile
😉
Kind regards,
Gift Peddie
Viewing 13 posts - 16 through 27 (of 27 total)
You must be logged in to reply to this topic. Login to reply