August 8, 2013 at 8:30 pm
Comments posted to this topic are about the item Would You Choose SQL Server?
August 8, 2013 at 9:51 pm
I have to say "it depends" on what the app is supposed to do (including "scale") and how specialized it needs to be. Having written a small and specialized "data engine", including my own sort routines, clustered and nonclustered indexes (non B-Tree), date/time functions, etc, etc, I might not even use a database engine. Of course, I'd also have to weigh out whether doing such a thing were cheaper than buying SQL Server and what the hidden cost of people hating proprietary databases might be. They might not buy the product unless it uses a recognized/established dtabase engine.
--Jeff Moden
Change is inevitable... Change for the better is not.
August 9, 2013 at 1:50 am
As Jeff says, it depends on the scale.
But if you are talking about a small business startup I'd say use AWS and one of the RDS options as a starting point.
The other thing is to look at your requirements. If you suddenly find that you need an RDBMS and capabilities that look like SSRS, SSAS and SSIS then I'd look at SQL Server. If you only need a small part of the SQL Server eco system then perhaps not.
I'd reitterate that the decision has nothing to do with the quality or lack of features within SQL Server. It is purely on getting an acceptable solution at an acceptable cost to start and grow the business. Remember that a high proportion of business startups fail in the first 18 months.
August 9, 2013 at 6:14 am
So, I'll sort of echo what's already been said...
If we're looking at just the data engine, I'd look at other products, too.
But as soon as the requirements start needing ETL and reporting tools, SQL Server gets a lot more attractive.
Thanks,
MKE Data Guy
August 9, 2013 at 6:15 am
Any decision as to which database to use depends on what the application will do today and for the forseeable future. Even if you are a SQL Server shop, it pays to analyze whether SQL Server is the correct platform. Sometimes the decision is made for you. If the vendor of a package that has already been purchased says the platform is MySQL on Linux then that's what it is.
August 9, 2013 at 11:28 am
Second David Poole's comments. I am looking at building small apps for researchers and AWS is probably how I'm going to go. Even using free SQL Express wasn't interesting to me.
But for a large enterprise MS SQL would get considered.
August 9, 2013 at 4:55 pm
I guess I have to agree with everything that people have said so far.
However, it's probably worth saying that while MSDE was not really a consideration for anything but trivial applications, SQL Express is much more useful. So there's a class of app for which SQL express is adequate, and it's certainly difficult to claim that the license for that is too expensive.
I think some of the changes in licensing recently have made SQL Server much less competitive that it used to be. If I understand the increased costs correctly, there are many cases where a couple of years ago I would have recommended SQL Server but today I would recommend Postgres.
Tom
August 9, 2013 at 9:32 pm
No I would not buy SQL Standard to build the app on the hardware you describe. I would by Developer Edition write the SQL code in ANSI standard and then load test it to see what hardware was required and see if I would have enough customers to make it profitable to run on Standard or needed to be ported to another SQL product.
However I completely agree that 2012's licensing was a major step in the wrong direction.
August 9, 2013 at 9:37 pm
Don't even start with the Express version. We started with SQL Standard, then had memory / CPU constraints and moved to SQL Enterprise, being the easier upgrade path. Now pricing is ridiculous and web scale is limited, so we're looking to get out from under the cost/limitations of it. Unfortunately; we have a lot of development, years of it, around MSSQL at this point and migration is difficult. If I had to do it over, I'd find another solution, there are plenty, one that web scaled and was either open source or cost effective at web scale. MSSQL is not!
August 10, 2013 at 5:03 pm
Like everything ... it depends. Generally if you are being paid for and paying for (ie have customers and are paying employees salaries) you have to manage some combination of what you are best able to use as a tool and what your customers demand. 8k is not a lot it is about 1 developer month worth of expense. So if you are building anything that you expect to use more than say 6 people months developing the SQL cost is relatively trivial.
If you are a small shop/just starting out try some combination of SQL Express, Azure, or getting hooked up with BizSpark. By the time you need to scale up you should have the revenue to do so. It is all what you as a developer and your customers can support in house. In my experience the expectation is that MySQL/PostGres will be a free/cheap offering and anything on Sybase/DB2/MS SQL etc (the enterprise dbs) will be at "professional" prices. Part of that is they have the built in expense of the database license but I think a part of that is customers mindset IMO too. Give me a computer department ticketing system running on MySQL and I'll assume it is FOSS or try to find a free one. Give my one running on SQL and I'll expect to spend $2000 a year and have someone to call that one time in 10 years I need customer support. If the problem it solves is worth the cost I don't care one way or another make the problem go away and I'll be happy. I suspect most customers are that way too (and the ones not willing to pay a profitable amount for the development cost of the software aren't really customers worth having are they?).
August 10, 2013 at 9:57 pm
I work for a SW company. We host the customer's data and they access it via RDP as well the customer doing self-hosting.
So when you're designing you app and making a DB selection, you have to look at what customer base you are aiming at. We have customers that are formal, large corporations in the Fortune 1000 crowd down to a small single standalone facility. The one is so small that the on-staff IT contact is the facility maintenance guy.
We were selling one app --call it App O -- that we absorbed the MS SQL licensing in the setup and monthly maint fees. The costs were such that we would have had to get over 800 users to continue development. We never got above 173 customers using it.
We also have one app that uses Interbase. It's a good solid app (App Ac). You don't even really need a server to run it, but it helps. But App Ac never had the same capabilities as App O, and the development was stifled because we were trying to sell App O.
I won't mention the other because it is a character based system from the 80's that is off a debased FoxBase corruption.
But if you are developing for an expected market of Fortune 500 instead of developing for mom & pop shops makes a difference.
----------------
Jim P.
A little bit of this and a little byte of that can cause bloatware.
August 11, 2013 at 9:15 am
Jim P. (8/10/2013)
But if you are developing for an expected market of Fortune 500 instead of developing for mom & pop shops makes a difference.
I'm not so sure it does in every case: we delivered to very big companies, but they were wanting to install the system on many sites, and tens or hundreds of processor licenses for SQL Server standard edition is big enough even for a big company to look hard at the cost.
Tom
August 11, 2013 at 11:43 am
I'd like to add something with regard to open source software.
1. community editions may be free and will get you started but once you get beyond a certain point you will want the commercial editions which are definitely NOT free.
2. If you need support then this is NOT free. In fact "professional services" can be very expensive.
3. If your open source choice is not something you would or could contribute something to then what is the benefit. I couldn't contribute to MySQL or PostGres code. I might add something to Hibernate or Solr at a push.
4. Open source projects seem to attract the masses. It used to be that only the big players could afford an MPP appliance and only they had "big data". Now that many people say they have "big data" projects such as Hadoop start to come to the fore.
5. Open source projects with some overriding governance seem to be the ones that survive. The rest fragments and die.
6. Very few people can afford to work for free. Businesses chasing free software from companies without a viable business plan are going to be left holding an ugly baby.
August 12, 2013 at 1:12 am
As usual the answer is "It depends .... "
What other software and how much of it is the business using - if you're a Microsoft shop as we are, you could find an enterprise agreement covering all your software licencing including Windows, Office and SQL Server is better value.
August 12, 2013 at 6:48 am
It depends.
We have small clients that can't afford SQL Server, DBA's and the infrastructure, etc... But their data can not be hosted on the cloud due to legal reasons. So do we absorb the cost and host their data and manage it? Yes, in some cases that works. But we could also put their data in OSS solutions and save them and us an large chunk of change.
I think the deciding factor is that today Microsoft really only cares about the enterprise clients feature wise and security wise. This opens the same door for solutions that Microsoft came through in the '90s when they started with SQL Server and cared about smaller organizations.
Viewing 15 posts - 1 through 15 (of 37 total)
You must be logged in to reply to this topic. Login to reply