August 29, 2010 at 3:05 pm
I know this is a subjective question, and a topic for someones masters thesis/dissertation should be a topic of interest to the student in question, but i have no idea whatsoever and am looking for some inspiration.
Im hopefully going to do a masters in computer science in the next year, part time. I need to do a thesis/dissertation that also should include some form of demonstration of the topic i have been investigating.
To give an idea of my interests, i work on a daily basis in Application Support and Administration, mainly administering sql server , debugging large highly nested sprocs, ensuring production environments stay up and running and reviewing func spec's, amongst other things.
Can anyone throw out a few topics in the database area that may be a decent idea for a dissertation ?
I would like to do something surrounding disaster recovery and planning, as i will be getting into this area at work soon.
I also spend a huge amount of time walking through sprocs that are thousands of lines long, and can be nested many levels deep, to identify bugs, and although i cant think of anything at present, im sure there are good topics for a dissertation in this area.
Hopefully Il get some inspiration from your suggestions!
August 29, 2010 at 6:50 pm
In terms of DR, perhaps something that examines the cost of data loss and the risk factors and maybe determines some formula for how much $$ should be spent?
For procs, I'm not sure. Is there some guideline about what the thesis should be about? Maybe an analysis of time spent on procs in debug v development and any correlation?
August 30, 2010 at 3:51 am
Steve Jones - Editor (8/29/2010)
In terms of DR, perhaps something that examines the cost of data loss and the risk factors and maybe determines some formula for how much $$ should be spent?For procs, I'm not sure. Is there some guideline about what the thesis should be about? Maybe an analysis of time spent on procs in debug v development and any correlation?
Both good ideas.
Is there some guideline about what the thesis should be about?
The course has'nt started yet, and im still waiting on the schedule to be sent to be sure i can do it around my current job, so i have not yet had the chance to converse with lecturers on topics, duration and level of effort required.
Im trying to get a bit of a head start in comming up with an idea, as for projects like this,finding an idea that i will enjoy researching is always the toughest part for me.
August 30, 2010 at 8:19 am
Post back here if you know more. I'll follow the thread and if I think of anything else, be happy to add ideas.
August 30, 2010 at 11:55 am
Thanks Steve, much appreciated. If i can think of any more potential areas of study, and when i get more info, il follow up.
August 31, 2010 at 2:06 am
I'd suggest that you first speak with the lecturers and/or your thesis advisor. They'll likely have their own ideas and requirements. The thesis will likely need to be original research (so not just summarising known information) and hence contribute to the body of academic knowledge. Plus you'll need to do a lot of reading to see what's already been done in the field and what you can build upon. (scholar.google.com is a great academic search engine)
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
September 1, 2010 at 9:04 am
to expand on Steve's idea about DR, it also depends on what kind of disaster you're preparing for and the size of the organization. most of the DR push is from consulting and hardware companies who have a financial interest in selling you services to set up a site.
unless you're the NYSE or some other large company that is going to survive a major disaster it makes very little sense to spend a lot of money on DR. if you work for a smaller company than there is a good chance it's going to go out of business if there is a major WMD attack or earthquake or whatever. in this case it probably makes sense to put some old hardware at your DR site to keep the auditors away and that's it
September 1, 2010 at 9:14 am
alen teplitsky (9/1/2010)
unless you're the NYSE or some other large company that is going to survive a major disaster it makes very little sense to spend a lot of money on DR. if you work for a smaller company than there is a good chance it's going to go out of business if there is a major WMD attack or earthquake or whatever. in this case it probably makes sense to put some old hardware at your DR site to keep the auditors away and that's it
Wow. I have to seriously disagree with everything in this statement. Without DR, there is no business continuity and DR counts for more than major events. It counts for power outages, hard drive crashes, network card glitches, etc.
The price of disaster recovery ranges from inexpensive to highly expensive. There are options for everyone, regardless of business size, and I don't recommend any business do what you just posted. That's just asking for trouble.
Getting back to topic, if you're going to do a thesis about SQL Server, than I recommend finding a topic most people haven't thought about covering (which other posters have mentioned), or a topic in which you can create a new solution that hasn't been talked about. What about case studies of business hardware / DB models? What about data warehouse versioning?
September 1, 2010 at 9:20 am
Brandie Tarvin (9/1/2010)
alen teplitsky (9/1/2010)
unless you're the NYSE or some other large company that is going to survive a major disaster it makes very little sense to spend a lot of money on DR. if you work for a smaller company than there is a good chance it's going to go out of business if there is a major WMD attack or earthquake or whatever. in this case it probably makes sense to put some old hardware at your DR site to keep the auditors away and that's itWow. I have to seriously disagree with everything in this statement. Without DR, there is no business continuity and DR counts for more than major events. It counts for power outages, hard drive crashes, network card glitches, etc.
The price of disaster recovery ranges from inexpensive to highly expensive. There are options for everyone, regardless of business size, and I don't recommend any business do what you just posted. That's just asking for trouble.
I second this -- emphatically. I think every organization needs to include DR in their planning.
And BTW, this comes from experience . . .
My previous employer had an office in the World Trade Center.
+--------------------------------------------------------------------------------------+
Check out my blog at https://pianorayk.wordpress.com/
September 1, 2010 at 9:25 am
i didn't say no DR, i'm against spending a gazillion $$$ on brand new hardware that will sit unused 99.9% of the time. for most of the events like a hard drive crash you can set up clustering for a lot cheaper than DR. and a lot of organizations will have redundant switching/routing at the primary site anyway. you can set up a good enough DR site by buying new servers every 3 years or so and sending the older ones to a DR site. chances are they will be good enough to operate in a DR scenario. i've been reading tech journalism for 10 years and 95% of the time it's rubbish and advertising posing as journalism. every few years there is some new fad they hype. the last time i fell for it was when they hyped IM. now it's cloud computing and DR.
if there is a nuclear suitcase bomb that goes off in NYC or a chemical agent or the 8.0 earthquake that people are predicting then there is a good chance of a lot of companies shutting down for good and the fancy DR hardware being sold off at auction
how many people will abandon their family and head on down to a DR site 200 miles away if there is a nuclear explosion anyway?
it's like cloud computing. I priced out EC2 and unless you're a light user or small company it costs less for a physical server. the reason is that for every computing resource Amazon sells you they have to have redundant DR for it. From HP's or Seagate's perspective it makes perfect sense. why sell a 146GB SAS hard drive to a customer when you can sell 5 times the raw storage to Amazon and have them resell you the same storage?
September 1, 2010 at 9:31 am
I'm with Alen. Probably a debate for another thread (or editorial 😉 ), but the amount spent on DR has to be in line with the risks.
September 1, 2010 at 9:35 am
After reading his clarification, it makes more sense. If planned properly, good DR shouldn't have unnecessary costs and wasted hardware. On that point, I will not argue.
(To the OP: hey, maybe this topic in and of itself could be good fodder for your dissertation! ;-))
+--------------------------------------------------------------------------------------+
Check out my blog at https://pianorayk.wordpress.com/
September 1, 2010 at 9:57 am
here is another idea i've thought about, the financial viability of cloud computing
there is a lot of value in virtualization but as all the time i think it's being hyped for profit. and in this case the hardware companies are behind it.
say you have a SQL server with 100GB worth of databases that are on your plain SAS hard drives and a server with a quad core CPU
the hype is to virtualize almost everything these days. Amazon is even pushing EC2 for almost every computing problem. and there was a story linked at http://slashdot.org a few months ago that Amazon was oversubscribing EC2 which got me thinking about this.
for the above server you will probably buy 6 hard drives. 2 for the OS, 2 for the data files and 2 for transaction logs. the cost will be around $1500 if you're buying 10000 RPM 146GB SAS hard drives. in HP's case they are re-branded Seagates most of the time with custom firmware.
now say you transfer this instance to EC2. I'm sure that in the case of Amazon they have enough redundant storage to fail over their entire infrastructure. in their case instead of buying 6 hard drives they have to buy at least 12. that's not even taking SAN's into account. if all of the EC2 storage is on SAN's with BCV's then they probably have to buy 18 hard drives to handle the same work load as your 6. much better for Seagate. and since you're taking data that sits on a server and putting it on a SAN you're also giving business to EMC, Netapp and probably a few other companies.
i haven't done a real financial analysis and it doesn't apply in all cases since a lot of people like EC2. especially small companies with uncertain cash flow. i've taken financial classes and know that it's very easy to go out of business while growing rapidly due to cash flow issues. but for larger companies with their own infrastructure I think EC2 doesn't make sense. and cloud computing is just a new label for what used to be called ASP's in the late 1990's or Application Service Providers.
September 1, 2010 at 10:03 am
Ray K (9/1/2010)
After reading his clarification, it makes more sense. If planned properly, good DR shouldn't have unnecessary costs and wasted hardware. On that point, I will not argue.
Agreed. And I apologize for misinterpreting your original post, Alen.
September 1, 2010 at 4:30 pm
Disaster Recovery would make a great thesis. As part of your research, find out the school's DR plan. See where it is weak and what improvements might be helpful. They might not even have a DR. And DR doesn't just have to be about computers, what if there was a major fire and the school lost several of its buildings. What is their DR for taking care of the students?
-SQLBill
Viewing 15 posts - 1 through 15 (of 15 total)
You must be logged in to reply to this topic. Login to reply