The current application is a kind of CRM application built upon MS Access. The application is for internal use. My job is to migrate it to ASP.NET web-based application. Now boss requires to keep Access as database and develop ASP.NET code against it.
My question is, is there any disadvantages of using Access as database in ASP.NET application? (e.g. optimistic concurrency issue?) Should I persuade boss to upgrade Access to MS-SQL?
Many thanks!
We've used Access as a backend for web sites with good success. It's cheap, can be used effectively by moderately skilled programmers, and you can store the MDB on a document server so it gets backed up.
Most IT people dislike Access, but from a business perspective, Access can be very valuable.
MS Access is notoriously unstable in multiuser environments. A WEB app is by definition heavily multi-user.
So IMHO leaving MS Access as underlying DB is a call for trouble. At least use SQL Express (it is free)
The problem you are going to face in upgrading from Access to MS-SQL is that there is a major cost investment for the application. If your company already has the infrastructure in place(licensing, hardware...) then you won't have such a hard fight to pursuade your boss.
As for a technical answer:
I'd say you need to let you boss know that access databases aren't ideal for concurrent usage which a web application suggests is the intended goal of the application. My view is that Access is for database information that a SMALL set of users will be simply using for small data entry and querying. NEVER use Access to build an enterprise-level solution.
If you are planning to upgrade a Microsoft Access database to SQL Server 2008, use the SQL Server Migration Assistant (SSMA) rather than the upsizing wizard built into MS
10+ tips for upsizing an Access database to SQL ServerAccess.
Your boss probably likes to do ad-hoc stuff with access / excel. If you move the DB to SQL Server Express you can use Access and it's linked table feature to let your boss keep doing his ad-hoc needs through Access while keeping the data in SQL Server Express. If you keep the linked tables named the same as the old physical ones all his reports and queries will should keep working.
I'm an Access promoter, but not for use on websites because Jet/ACE is not threadsafe (though Michael Kaplan once said that is is threadsafe if you access it via ADO/OLEDB; I don't quite understand how a database abstraction layer can wash away a characteristic of the underlying database engine it's calling, but if MichKa said, it's 99% likely to be true).
Now, the exceptions would be if you're using it for prototyping something that will use a different database, or if it's read-only, or is read-write but will only ever have a very small number of users.
Michael Kaplan's website, trigeminal.com, used to use a Jet database as the back end (it may still -- I don't know that MichKa ever changed it), and when that was his main website he reported getting 100K hits a day. But it's a read-only site, so fits my restrictions.
There are so many different alternatives and they are mostly easy to use that I just don't see the point of trying to use Jet/ACE as back end for a website. I'd never do it myself (all the websites I'm responsible for use MySQL).
Simply put, go with MSSQL. Express edition is free, and will give you everything you need to migrate away from Access. These articles are talking about Access applications specifically, but the same issues will plague you.
http://resources.zdnet.co.uk/articles/features/0,1000002000,39285074,00.htm
https://web.archive.org/web/1/http://techrepublic%2ecom%2ecom/5208-6230-0.html?forumID=102&threadID=205509&messageID=2136367
Related
Back in 2008 I remember running into a 3rd party tool that did 99.9% of the work of implementing task/role security. From what I recall it was simply an extension of what was built into .Net. You ran a script to add all the needed tables to your SQL Server then used the UI to go in and define all the tasks and then grouped the tasks into roles. The assigned users to a role. The power was that different roles could have the same task and it was all configured by the UI tool that came with this system. I also thought it was on CodePlex, but I don't recall the name. All I recall it was Microsoft's name with like Sql or something added to the name of it.
At the time, 2008 (VS2008 days), I was told by co-workers that Microsoft was slowing consuming the whole system into .net.
Anyone have any idea what the name of the thing was I am thinking of? Is it part of Microsoft .Net now?
Are you perhaps thinking of NetSqlAzMan, which uses a similar authorization model to Microsoft AzMan? Or perhaps you're thinking of AzMan itself, which has a UI and allows storage in SQL Server?
I have a web site I developed for displaying the results of some data analysis work I did. It relied on ASP.NET for the front end and connected to a MySQL back end utilising Entity Framework and LINQ extensively.
I chose MySQL because I personally have used it in the past and like the database, but this resulted in some serious issues when I had to deploy it to a hosting provider (incompatible connectors, access rights, etc.)
I am now getting ready to redevelop and expand the site and I am looking for some advice to avoid the issues I had last time.
The new DB has to serve two roles. The first is to be a data provider for the charts that are the output of the analysis work. These tables are straightforward, almost flat files, with 10 tables. One table has roughly 200k rows of data the rest have aprox 1200 rows of data each. There are little references or queries between the DB tables, but there are a few. This data is updated periodically by a back end process and does not need to be added to or edited by the user.
The second role of the DB would be as a basic persistent store for a standard user management system. It would need to manage data for adding/ removing clients, user names, passwords, access rights. etc. No financial data or super secure data is involved.
What database approach would you recommend that would give me easy deployment and management at a web host and still allow me to use both Entity Framework and LINQ effectively.
Second, what tools/frameworks should I consider as I rewrite this system. It is very graphical and data focused. Presentation of charts and information is the key factor in this site. Are there any new technologies or frameworks that would add specific value to what I am doing?
A few notes. I am a one man shop and I maintain the entire system myself so I am less worried about enterprise level frameworks than other people. My focus is on the easy development and deployment of the site. Maintainability is also a key factor.
I am also an experienced C# developer, but new to ASP.NET and the web side of things. The first version of this site was a big learning experience. It was good, but I wasted an enormous amount of time on just understanding new technologies and approaches. I am very open to learning, but I can't afford the time to get my head around a complete paradigm shift.
I am looking forward to your thoughts, thanks.
Doug
The natural choice would be SQL Server. I'd guess by your description that you are way under the maximum space limit of the SQL Server Express edition. I of course supports Entity Framework and the drivers are part of the .NET Framework, so no problem with third party assemblies here.
This will also open up the possibility to host your app in the cloud (Azure) later on, because SQL Azure in fact is a Microsoft SQL Server, so there is no overhead in supporting that.
Regarding user management - ASP.NET has this all build in (Membership, Role and Profile provider) and also a SQL Provider for which default tables are available. So you don't have to design your tables by yourself and it runs very naturally on SQL Server.
What would be the pros and cons of taking a large (16000 LOC) existing Access 2007/Forms application and porting it to SQL Server 2005, while keeping the Access forms on the front end, versus porting the front end to a .NET technology, say ASP.NET 4.0 or MVC3? The backend is definitely getting ported to SQL Server, with large schema modifications, so this question is about whether to use the existing Access Forms front end or not. I don't take throwing away existing working code lightly! On the other hand, the schema will be changing a lot.
If you had a lot of experience with ASP.NET/Oracle but not with Access or SQL Server, how might this change your answer?
Thanks,
Mike
There is nothing wrong with Access as a front-end to an RDBMS whether SQL Server or another
Porting the front-end but leaving an Access back-end would seem a bit weird: why bother? Access is a good presentation tool but mediocre DBMS. Why keep the mediocre?
Also, only change one component at once no matter what you do: data, then front end. Don't do a big bang approach and change the lot
Actually, comparing going with web application (ASP.NET) and stand-alone (Access) is not truly "apples-to-apples" comparison. It highly depends on your requirements.
Sure, going with web application will require much more developer's efforts comparing to building forms/reports in Access. But, the pay-off will be felt in the mid to long term.
You should also consider scalability, deployment, and availability issues when deciding to go/stay with Access.
I have a situation where information about a user is stored in the web application cache and when that information is updated in one application - I want to notify the other applications (running on the same machine) that the data should be removed from it's cache so it can be refreshed. Basically I need to keep cached data in sync across multiple asp.net applications.
I have started down the path of using a central web service to help coordinate the notifcations but it is turning out to be more complex than I think it needs to be.
Is there a way that one asp.net application can easily reach across to another on the same box to clear an item from the cache?
Is there a better way to achieve shared cached information than using the application cache?
I really want to create a way for apps to communicate in a loosely coupled way - I looked at nservice bus but the dependency on MSMQ scared me away - my client has had bad experiences with MSMQ and does not want to support an app that requires it.
Suggestions?
Michael
I agree with Hogan. Best is to use a shared database. I want to add to that that, when using SQL Server, you can use SQL Cache Dependency. This SQL Server mechanism allows notifications to applications in such a way that used caches can be invalided directly after a change is made to the data.
A shared database is probably going to cause you the least pain.
Edit
Note: ASP.NET allows you to make "cache clearing" triggers on SQL server changes. Should be a quick search in the cache examples on MSDN to find some examples. Thus when the user info stored in the cache changes in the DB the local cache copy will clear and be re-loaded from the DB.
There are commercial distributed caches available for .net other than Microsoft Velocity - NCache, Coherence, etc.
How about Velocity? It's a distributed cache that works between servers as well as between applications. It has PowerShell management and all sorts of documentation to get you going faster and be far more maintainable in the long-term.
What about COM/DCOM, using namespace System.Runtime.Remoting
The default asp.net membership provider uses a .mdf sql server database file in the app_code database.
How scalable is this in terms of calling a flat file database instead of running it in a standard sql environment?
Is this recommended only for small/medium traffic sites?
It's a reasonable trade off for any site that can run on one server. It's fairly reasonable for small to medium traffic sites.
When you grow to a point of a web farm, then you'll be better off with a separate server. Also, depending on how database dependent your application is, you may find better performance handing off SQL queries to a totally different server/processor to handle the database side.
I wouldn't recommend this for anything but a "learning" project.
For any real application, regardless of size, you don't know what type of "next feature" you will add. You want to have a real independent database in which you can delegate functionality to, in which you can set jobs to run independently, sit on a different HD, maybe splitting it into a different VM?
You can use SQL Express and still be "free', and it is better to do this seperation before the site grows and the DB is harder to move.