Make a Move or Wait and See
-
Today we have two rather contrasting articles in the newsletter. One is an Upgrade Guide for 2005, which is the type of article we tend to run. The other is actually a piece by more, maybe a rant, but it advises that you don't upgrade to SQL Server 2005.
Don't get me wrong, SQL Server 2005 is a great product, but my feeling is that you should wait for SQL Server 2008 is you haven't upgraded. Reasons are in the article, so I won't go into them here.
However the decision to upgrade any software, whether it's an application, or a base platform, is one that many of us face regularly. Actually most of us get told we're going to upgrade without us providing much input or influence on the decision, but we still deal with it.
Why do we upgrade? I know earlier in my career I often pushed to upgrades, trying to be on the bleeding edge and trying out new technology. It was cool for me as an IT guy or developer to get to work on new technology. However as I've grown older, I'm less likely to upgrade. I like the stability and comfort I have with the products I've gotten used to. I also think that less and less of the newer features have a big impact on productivity. Something you might argue wasn't the case a decade ago.
I do know that the software industry needs to continue to build better products and having maintenance and patches for older software doesn't pay the bills. Sometimes it makes more sense to just correct a lot of things when you're adding new features and just release a new version. Whether this is the best way to do business or not is something to debate, but it seems that's the current state of affairs.
However in many big software applications, we see third party shops providing support and even maintenance patches. Heck we even see patches for Windows from third parties sometimes. I'd like to think there's enough demand for someone to start a business providing third party support, but maybe we really don't want it. If you could still get patches and support for SQL Server 6.5, would it make sense to still run your business on it?
I don't think I'd do it, but if I could get SQL Server 2000 support for another 5 years I might stick with it.
-
I am stilling struggling with SQL Server 2005 and the service pack is still keeping coming, definitely SQL server 2008 will not be considered in the next year or so.
However migrating from SQL Server 2000 to SQL Server 2005 can take some of the new features espeically building the BI and using the reporting service, this is the only reason I start getting my foot wet.
-
I was reading all the feature blurbs for SQL Server 2008, and it got me to thinking that if I looked at the gap between the release of 2005 and its Service Packs, the degree of change in those SP releases was about the same as the 2008 leap. It then begs the question if this is actually a new product at all, or basically just 2005 SP3 or SP4. If the latter, then it makes one wonder some. Of course, I'm not being generous even suggesting that, but its a realistic view that I believe must be affecting adoption at some businesses. And if it is a new product, it begs the question why such a massive shift when they proclaimed the 2005 platform would be the one they would build upon?
But most disturbing to me lately, the past five years or so, is that when you look at the lock-step progression of the .NET frameworks and related products, what you see is instability built upon instability. The lack of support in VS 2005, for example, to code to .NET 1.1 was a kick in the chops to anyone who has to support old systems that are not yet upgraded. The version stamping of the Vista extensions as .NET 3.0 was a sad effort that smacked of spin; and the fact that OS supports so many of the development platforms so poorly is disheartening. The way the CLR was integrated into SQL Server 2005 was sloppy, and you can see it in how many peculiar bugs affected the new GUI elements. Regardless of the fact they solved many of them fast, the fact is they existed because of the integration of a maving target inside another moving target. And the feature-spread in the new management tools, regardless of what MS says, was affected by the awkwardness of implementation brought on by the reliance on .NET. Ultimately, the point I'm getting at is that the integration of these technologies has created a situation where the failure of one can have a dramatic impact on another. Just try managing your development databases in the GUI when the framework is messed up and you quickly come to hate the fact they are co-dependent.
So, as I read the plan to make February 2008 a point to release a new Visual Studio, a new Windows Server and a new SQL Server my first reaction was to shake my head sadly. In our shop VS 2005 is still barely pulling its weight, and while we migrated to SQl Server 2005, giving us a trio of upgrades early 2008 means making an awkward choice, because my gut feel is that if you go with the database engine upgrade, you will require the new VS to get the value from its features, and to have the engine run at its best you'll need OS support built into the new server, etc. So, again, it feels like MS is asking us to trust them to have the integration right at the first drop, and that seems more risky than ever (as much as I love the intent of VS 2005, it was released before it was baked, and that severely affected productivity).
All my rambling aside (its been one of those days), MS needs to start to consider its long-term revenue generation, and recognise it is based in time to implementation of enterprise solutions. The products are not being allowed to shine and show their value propositions before we are saddled with new lock-step releases, making adoption harder to sell.
Of course, this is just an opinion.
-
Frank,
Great points and to a large extent I agree. It seems to me that MS is really built to force half the installations to upgrade to every other release, so each "set" of releases (i.e. SQL2K5 and SQL2k8) will get everyone to upgrade between them. To met VS2005 probably gets cleaned up and VS 2008 is the mature version.
I thought .NET was a great idea, but I was as skeptical of it as I was in the early days of COM. The forcing of you to upgrade by not supporting previous versions kind of ruins .NET and gets you back into the .DLL versioning issues. You don't worry as much, but you have to now support multiple .NET versions, which can be a problem for some machines and software on the desktop.
-
We're already in the process of migrating to 2005. Although we're very early in the process I think we're going to proceed. My experience is that application vendors are always behind the curve with support for new platforms, especially in the DBMS area. Also, I never do an in-place upgrade of a SQL Server; instead, I build a new server with the latest version of SQL Server and move databases one at a time as applications add support for the new version. I am also uncomfortable with RTM releases of software so I will likely wait until 2009 before tackling SQL 2008.
[font="Tahoma"]Bryant E. Byrd, BSSE MCDBA MCAD[/font]
Business Intelligence Administrator
MSBI Administration Blog -
Bryant, good luck on the move.
I still think it's worth waiting to move (though not to test). The RTM was announced yesterday and I see 2008 as a SS2K5 R2, not a new set of bits.
However I'd definitely be interested in hearing about issues upgrading.
-
Change is good; stability is better.
It seems most of the upgrades are just change, result in a new learning curve, and introduce instability at first; and the SW industry (MS + Others) just doesn't seem to get it. One example is Vista.
I'd love to know what the true Vista adoption rate has been (not the Vista licenses bundled with new PC sales, but the Vista penetration into the currently installed base of BUSINESS). My guess is that it is relatively low - I know of no company which has implemented VISTA wholesale - because if you really boil it down, the whiz-bang features of VISTA don't significantly impact productivity or add compelling features. And I'd also propose that the change in the Vista UI (I'm sorry, the "User Experience") will actually result in lost productivity in the short term once Vista is adopted. Plus the business IT powers to be understand that as long as WinXP systems are properly patched with good AV software, the security flaws advertised as fixed in Vista can be overcome in XP.
With the midnight introduction of Win95, MS either fell into or planned a new strategy of compelling business upgrades: let the home users demand it from thier companies. They seemed to be pushing this with Vista with all of the "MS Events" through April or May of this year. But the strategy no longer seems to be working. So it appears thier strategy has changed: put out an upgrade for everything making the products so inter-dependent, that the advantage of certain features won't be gained as long as one product isn't upgraded. Time will tell if that strategy works.
In contrast certain platforms persist because they just do the task intended and require little maintenance (how many of us know someone who is still running Windows 98?). Since most of the industry doesn't exist for itself, but serves a company selling a product or service - upgrades become the mechanism to make compelling features available to applications. And most of the available upgrades offer so few real advantages there isn't a compelling reason to move.
Change is good; stability is better.
-
I wonder is it worth waiting for 2008 ver and then go for MCST and MCITP Certification exams or go now and then pay for extra upgrade cert exams.
I don't understand this short time between versions, who acctualy benefit from this???
-------------------------------------------------------------
"It takes 15 minutes to learn the game and a lifetime to master"
"Share your knowledge. It's a way to achieve immortality." -
We have one application, which is the backbone accounting system of the company that was built around SQL2000. Upgrading the SQL version gains no benefit to us and if past experience has shown, would probably break the accounting software package.
Thus, we have no incentive to upgrade a perfectly stable running SQL system until we upgrade the accounting package.
I would suspect others are in the same boat. Can't upgrade unless the application manufacture supports it.
I would also suspect that M$ will make the upgrade path more difficult the further apart the versions e.g. 2000 to 2011 in one leap.
Look what they have done to Exchange and W2K3 server with only a two version spread between products.
-
My organisation is in the midst of upgrading a few identified databases to SQL2005 more for business reasons. I do agree with Steve that from a business and technical perspective, it does make sense to wait for SQL2008.
But on the other hand, new functionalities such as row versioning, better SQLOS (better SQL scheduler), etc will definitely give an advantage to get to 2005 for large and highly concurrent DBs (E.g. SAP is already certified for SQL2005). I might be wrong, but I wouldnt want to wait for users to complain for the next year or so and ended up with an executive decision to scrap the app. For all the management care is, the system works smoothly and less grouse with users.
We've got policy to put new DBs on SQL2005, but not on SQL2K. I'm hoping M$ will provide a smooth tool to migrate to SQL2008 when the product is launched.
Our organisation is very large and I've been having discussion directly with few M$ managers. They were interested in what we think and how we use SQL2005 in production. From what we've been talking to far, they were still in the "requirement gathering" phase for SQL2005, asking feedback on what functionality, etc desired/useful to be included in SQL2008. I reckon M$ has incorporated this process since the past few SQL releases.
Simon Liew
Microsoft Certified Master: SQL Server 2008 -
I have used VS2005 for development both with Oracle and SQL Server without problems, strange the integration with Oracle both 9i and 10g was smoother than SQL Server in team development. Although I was also using SSRS when I was using SQL Server but not with I Oracle but the developer across was using SSRS with Oracle. I must add developers just write code no deployment or testing, there are separate teams for that and we don't know each other.
Kind regards,
Gift Peddie -
All of our systems are still on 2000 (with one major app still on SS2K+SP3a
)I was preparing myself to convert our systems over in the next 3-to-4 months (if I can get the people who sort out the licences to extract their thumbs from their collective backsides!
) with the intention to not even look at 2008 until closer to the end of 2008 or even 2009 - but now I'm not so sure.....With 2008 being touted as a very similar interface as 2005 (not looked yet, myself), I think I am in agreement with the previous poster who said that it may well appear to be 2005+SP3 or SP4.
I think I will still push to go '05 before the end of this year so that I will at least have a grounding in SSIS and various other features before going for '08.
A lack of planning on your part does not constitute an emergency on mine.
Viewing 13 posts - 1 through 12 (of 12 total)
You must be logged in to reply to this topic. Login to reply