Planning on using Notification Services? Don't bother…

Tonight we saw the release of the July CTP (odd name for something released late July 31st, but that's another story) of Katmai. With it comes the following blurb, deep in the fifth section of the readme file:

SQL Server Notification Services will not be included as a component of SQL Server 2008, but will continue to be supported as part of the SQL Server 2005 product support life-cycle. Moving forward, support for key notification scenarios will be incorporated into SQL Server Reporting Services. Existing Reporting Services functionality, such as data driven subscriptions, addresses some of the notification requirements. Features to support additional notification scenarios may be expected in future releases.

What strikes most of us as being, well, odd to say the least, is that there has been no formal announcement of this, nor is one planned.

Aaron Bertrand

I am a passionate technologist with industry experience dating back to Classic ASP and SQL Server 6.5. I am a long-time Microsoft MVP, write at Simple Talk, SQLPerformance, and MSSQLTips, and have had the honor of speaking at more conferences than I can remember. In non-tech life, I am a father of two, a huge hockey and football fan, and my pronouns are he/him. If I've helped you out, consider thanking me with a coffee. :-)

23 Responses

  1. Ubaid says:

    SSNS should not be dropped from the SQL server 2008, since there are people who are using SSNS as the base line for their event handling.
    I agree that it is hard to work on it in the begining and the users of SSNS are not in larger number, but these are not good reasons to not to have that in further editions.

  2. Gerard says:

    I second Robert!
     The company where I work need such a solution… I won't start a NS project if it's dropped in 2008!
    [email protected]!
    Ty!

  3. Robert says:

    I second Michael's query re white papers you'd (Tim McCurdy) be willing to share.
    The company that I work for has the need for such a product but is reluctant to use SSNS 2005 as it goes away with SQL Server 2008. We'll likely build our own or possibly, though not probably, run an instance of SQL Server 2005 forever.  [email protected]
    Regards,
    Robert

  4. Michael de Groot says:

    Similar to "newbe's" request, can anyone point to other 3rd party enterprise systems that support alerting?  Tim, was your solution custom built?  How has it performed/scaled?  Do you have any white papers or architectural documents about the approach that you could share?  [email protected]
    Thanks,
    Michael

  5. newbe says:

    Can anyone tell me the main ingredients to doing alerts without NS. I was planning on starting a SQL NS project for my familys small business and now I guess I will have to do it by hand. Any ideas on how to get started? [email protected]

  6. Tim McCurdy says:

    This is great news!  I hate NS!  We basically created a complete NS rewrite in 3 months and it doesn't require separate databases and runs out of 4 tables – not tables and tables of "chronicles" that now need to be managed and backed up.  It also works with the Service Broker so that databases in other servers can take advantage of one "hosted" event / notification database throughout the enterprise.

  7. Steve Rezhener says:

    I'm not sure that MS totally killed NS.
    "Katmai" July CTP still includes templates for NS and SQL Server 2008 Books Online still includes "How-to"s. However, there is no visible entity that called Notification Services in SSMS object explorer anymore.

  8. Howard Ryan says:

    It comes as no surprise to us that 2008 will not be supporting notification services.  Several years ago we developed our product line of alerting applications and gave some long and serious thought into using SQL Notification services.  
    Besides the fact that the desktop alert component itself was rather bland/limited, at the time it just made no sense to create a core component of the business plan around such a new MS widget.
    We did it the old fashioned way.  We wrote every line of code for our client server alerting system.  The cost was simple.  We would relinquish a near term competitive advantage to other alerting companies who decided to build their alert platform around NS.  While a look at this MS page shows only a few,
    http://www.microsoft.com/sql/technologies/notification/partners/appsoftware.mspx
    There are many more companies who must now consider a total rewrite of their alert system.
    Quite frankly, looking forward at MS's new family of Communications servers,
    http://msdn2.microsoft.com/en-us/office/aa905499.aspx
    I am not sure that things are changing much.
    "Welcome to the Microsoft Office Live Communications Server developers' home on MSDN. Live Communications Server 2005 provides your business with an enterprise-ready instant messaging, presence awareness, and an extensible platform that connects people, information, and business processes—enabling better decisions faster."
    Remember this?
    MSN Buys MessageCast for Real-Time Alerts
    http://www.eweek.com/c/a/Messaging-and-Collaboration/MSN-Buys-MessageCast-for-RealTime-Alerts/
    MessageCast was a serious product on the up.  They sold out and poof.  The whole fiasco was a joke.
    As it pertains to alerting platforms, the good call just may be having a architecture that is somewhat autonomous ( i.e.  Ok, perhaps Windows OS, .Net Framework and SQL Server as prerequisites is ok) to MS embedded components and "Live Server" licenses.  Seems like just a better bet hands down all day long.
    Our customers report to us, many of the alerts sent out are to inform users of downed MS services such as Exchange.
    So, the alert system if designed with an organic self-written architecture (the hard way) may just be the heads up alert your customer needs to see.
    That’s our two cents anyway.

  9. Howard Ryan says:

    It comes as no surprise to us that 2008 will not be supporting notification services.  Several years ago we developed our product line of alerting applications and gave some long and serious thought into using SQL Notification services.  
    Besides the fact that the desktop alert component itself was rather bland/limited, at the time it just made no sense to create a core component of the business plan around such a new MS widget.
    We did it the old fashioned way.  We wrote every line of code for our client server alerting system.  The cost was simple.  We would relinquish a near term competitive advantage to other alerting companies who decided to build their alert platform around NS.  While a look at this MS page shows only a few,
    http://www.microsoft.com/sql/technologies/notification/partners/appsoftware.mspx
    There are many more companies who must now consider a total rewrite of their alert system.
    Quite frankly, looking forward at MS's new family of Communications servers,
    http://msdn2.microsoft.com/en-us/office/aa905499.aspx
    I am not sure that things are changing much.
    "Welcome to the Microsoft Office Live Communications Server developers' home on MSDN. Live Communications Server 2005 provides your business with an enterprise-ready instant messaging, presence awareness, and an extensible platform that connects people, information, and business processes—enabling better decisions faster."
    Remember this?
    MSN Buys MessageCast for Real-Time Alerts
    http://www.eweek.com/c/a/Messaging-and-Collaboration/MSN-Buys-MessageCast-for-RealTime-Alerts/
    MessageCast was a serious product on the up.  They sold out and poof.  The whole fiasco was a joke.
    As it pertains to alerting platforms, the good call just may be having a architecture that is somewhat autonomous ( i.e.  Ok, perhaps Windows OS, .Net Framework and SQL Server as prerequisites is ok) to MS embedded components and "Live Server" licenses.  Seems like just a better bet hands down all day long.
    Our customers report to us, many of the alerts sent out are to inform users of downed MS services such as Exchange.
    So, the alert system if designed with an organic self-written architecture (the hard way) may just be the heads up alert your customer needs to see.
    That’s our two cents anyway.

  10. John Stephens says:

    We have a large application in production use that uses Notifications Services. In fact, we are currently developing even more enhancements to be released in the next month.
    The complete lack of roadmap by Microsoft is not acceptable. I don't understand why they took a stand alone add-on, integrated it, then killed it. I'm fine if they don't want to enhance it, but it needs to be made availabe as a separate add-on again.

  11. natasha says:

    thank you for the post, so if we want to use sql server 2008 and we need the functionalities of NS, what are we supposed to do ? is there a way we can install NS component from 2005 on a SQL server 2008 instant ?

  12. Ollie Riches says:

    So does this mean there will be no support for a similar API iun 2008 or are they rolling the functionality into RS – I personally architected a solution to use NS for an application, produced a spike (prototype) and got it all working perfectly in less a week, unfortunately the project got canned.
    Are we going to have to go back to the 2000 approach – triggers and extended stored procedures…

  13. Abu Pinhus says:

    I suggested to use it (as architect) , but  never get to do it.  Developers  spend a week (billable!) try to get thru documentation and gave up. So we used triggers and C# instead.

  14. fafnir says:

    …This is unbelievable: no more NS !!!… RS instead!!!..
    At my previous job, I created a NS 2.0 application which had absolutely NOTHING to do with Reporting Services.
    Of course, if and when NS becomes a part of RS, folks will use RS for the same purpose…
    but what's the point?
    Does MS have a good reason for this kind of change?
    NS was available only for SQL Server Dev/Standard/Enterprise editions, so it was not exactly "free stuff". Therefore, there's going to be no financial gain here… But I might be wrong about that, of course.
    The only explanation for this change that I can think of is a goal of making things a bit simpler for developers: once you've learned RS, it means you now know NS as well, right?
    By the way, I disagree with Adam about poor documentation:
    1. BOL have a good tutorial.
    2. I bought a book by Shyam Pather (who's the Project Manager for NS at MS) about NS 2.0 (NS for SQL Server 2000), did almost all the samples, and it took me only 15 days from the moment I started reading it till I had a fully functional app, including a custom content formatter. And I was the only developer on the project. …well, I read this book at work, on a train to/from work, at home, and of course in a bathroom… 🙂

  15. Davide Mauri says:

    Hi Adam
    I think that could be a really cool project! The only "problem" is time 🙂 … I think that a project like that would take at least a month (solar month) of development (just to realase something usable)
    I should also emulate current NS API so that NS can be taken away and substituted with that project seamlessy (hopefully)
    Anyway, I'll surely take part to that project if we choose to start it. 🙂

  16. cinahcaM madA says:

    It's a great idea to do an open source project, but I think you're too late 🙁
    More interesting idea, for me at least, would be to start a new open source project which would be a replacement for SSNS, based on SSB+SQLXML+SQLCLR+Database Mail–and which would also include a UI, perhaps as a plugin for SSMS.  What do you think?

  17. Davide Mauri says:

    Hi Adam
    In fact for both customer the first thing I did was a GUI to allow them to manage Subscribers, Subscriptions and everything else. As an Open Source project we can contribute to do something completely re-usable (I've already have a little prototype)

  18. cinahcaM madA says:

    I agree, it's a good product ONCE you get the hang of it.  And that's really the problem.  I believe that the reason its adoption rate was so poor was that its documentation is incredibly bad–terrible, really–and the product itself is not especially well thought out.  It is very awkward to work with and somewhat nonsensical in places.  A great UI wrapper on top of the XML and a rework of the documentation would have created an easy to use product that people would be much more likely to pick up and run with… I think that would have been a much better choice than dropping it.

  19. Davide Mauri says:

    I have two project in production.
    Though NS is really hard to work with at the beginning, it allows to create enteprise-notification solution that works really well.
    A big customer here in Italy is using NS to handle personalized event handling for more than 1000 employees.

  20. cinahcaM madA says:

    I've done two projects with it, although one of them may have never gone to production so I guess that one might not count 🙂

  21. Aaron Bertrand says:

    There isn't a ton of traffic in newsgroups like microsoft.public.sqlserver.notificationsvcs and the forum http://forums.microsoft.com/MSDN/ShowForum.aspx?ForumID=97&SiteID=1
    …but it isn't zero, either.  🙂

  22. Denis Gobo says:

    I really wonder how many people actually use Notification Services. I don't and I haven't met anyone yet that does either

  23. Davide Mauri says:

    WTF!
    But if they don't want to support NS anyway why don't release source code on CodePlex a make it a "open" project?