April 11, 2012 at 9:26 pm
Comments posted to this topic are about the item Transfer Times in the Cloud
April 12, 2012 at 12:44 am
Now your getting somewhere!
Earlier, I was under the distinct impression that you liked the cloud services. But because an idea might be good it does not mean the concept created based on that idea is good.
April 12, 2012 at 6:06 am
Well, there sure is historical record to look at, if it can be found. IBM ran its Service Bureau (which amounted to a private cloud) for decades. How did its clients get on when trying to get out? Yes, cloud is insidious. Yet another lemming road to ruin. It will be interesting to see whether experience and intelligence will win out. CIO/CTO types do tend to be sheep.
April 12, 2012 at 6:16 am
In reality aren't you concerned about ALL of the things you mentioned (Performance, Reliability, Security, Legal, Data independence). Why are people and companies so eager to give away their data?
April 12, 2012 at 6:48 am
Another point to consider - what if the company hosting your data goes bankrupt? The receivers will likely turn off the servers abruptly, without giving you any time to extract your data. While you might eventually be able to get your data back, it could take months or even years.
April 12, 2012 at 6:56 am
Oddly, this is also one of the argument used against SQL databases: what if XYZ Database Corp. goes under? Let's just continue writing COBOL/java/VB with our own fantastic I/O modules. But, they'll happily give over the data to an outside force, which is what really matters, after all (the code??, not so much).
April 12, 2012 at 8:07 am
I don't like the idea of it at all for the specific reasons you state. You give up complete control going to the cloud. I just don't see the great benefit to this. How can you be 100% satisfied that they are backing up your data and it can be recovered any day of the week?
April 12, 2012 at 8:11 am
IceDread (4/12/2012)
Now your getting somewhere!Earlier, I was under the distinct impression that you liked the cloud services. But because an idea might be good it does not mean the concept created based on that idea is good.
I like the idea for a few reasons, but so far the implementation has been "meh" to me.
April 12, 2012 at 8:12 am
richardd (4/12/2012)
Another point to consider - what if the company hosting your data goes bankrupt? The receivers will likely turn off the servers abruptly, without giving you any time to extract your data. While you might eventually be able to get your data back, it could take months or even years.
Write an escrow of the data into your contract. Require a regular copy that is available under any legal filing. We've done this in the past with source code from smaller vendors.
April 12, 2012 at 11:41 am
I have a slightly different viewpoint on the subject. I work for a company providing cloud hosting services to our customers.
We set up our customer environments such that they get visibility into the entire VM with full Win and SQL admin rights, no different than if they were running on a physical server. They're free to import/export data, copy off their backup files, replicate/logship/mirror to a DR target, or any other activity that keeps their data safe according to their unique business requirements. And, if the customer leaves they're free to take their data with them.
As with any cloud, sure there are tradeoffs with having your data living in a shared environment. It's a deal-breaker for some companies, a non-issue for others. Being a DBA myself, I certainly understand the aversion around placing your data "out there somewhere", but we do run our own internal databases on our own cloud environment without any problems.
One benefit with cloud that is not always obvious to customers is that we can host hundreds of virtual servers in the same cabinet space as a few dozen physical, and that's a cost savings we can pass on. That makes cloud very attractive.
April 12, 2012 at 4:18 pm
I don't trust the Cloud. I would only use it for Encrypted Backup Storage if I had no other choice. I see data hijacking/kidnapping as a real threat as this moves forward.
April 13, 2012 at 2:51 pm
I don't trust the cloud
Heh. I can think of a few organisations whose data would be safer stored in a professionally run cloud environment than in their own local mess. I'd suggest that very small, dare I say "backyard" organisations, are often not going to have the breadth of expertise needed to manage large and growing amounts of data. The cloud can provide the economies of scale to overcome this (to some extent).
Having said that, certain people in my organisation took it upon themselves to purchase software services where our data (at least I would have thought it was OUR data) is hosted externally, and it has been one problem after another trying to set up a fast efficient ETL to get timely data for reporting... especially since the external providers in these cases had never even heard of ETL.
...One of the symptoms of an approaching nervous breakdown is the belief that ones work is terribly important.... Bertrand Russell
April 13, 2012 at 7:47 pm
Randy Rabin (4/12/2012)
One benefit with cloud that is not always obvious to customers is that we can host hundreds of virtual servers in the same cabinet space as a few dozen physical, and that's a cost savings we can pass on. That makes cloud very attractive.
So, when the federal police for whatever country the physical servers happen to be in walk into the "cloud" provider's space, and walk out with all physical devices used by client X... what about all those other clients, and their data?
For any clients with legal and regulatory privacy restriction, I'd also worry about thin provisioned disks that can be compacted; my database's tempdb grows, so does my VM, so does its VMFS allocation. When my database and tempdb and VM uses less space later, during the compaction, is all the space overwritten, or can the next user that gets it allocated retrieve any of my data?
As far as the original article goes, getting your data out... after it's grown over N years to Y times the original size... is potentially a very serious challenge, and the math should be (and won't be) rechecked every year, and the plan should be (and won't be) actually tested every year or two. That's true for all products - "how do I move to a competitor's platform later" is a question hardly anyone asks for hardly any product.
Viewing 13 posts - 1 through 12 (of 12 total)
You must be logged in to reply to this topic. Login to reply