Voice of the DBA

Category: Editorial

  • The Speed of Azure

    We should be able to deploy a new SQL instance as easy as this.
    We should be able to deploy a new SQL instance as easy as this.

    As database administrators we seem to be slow to embrace new technologies and paradigms. There was a lack of enthhusiasm from DBAs for SANs years ago, and virtual machines more recently, at a time when many other technology professioanls were embracing these ideas. Even today there is resistance from some people, and sometimes with good reason. SQL Servers are not like other servers and have much different hardware requirements. Too often the virtual machine and storage administrators do not appreciate that SQL Servers need different architectures.

    Lately the cloud services push is seeing lots of resistance from DBAs. Various vendors and media hype the idea and potential savings, which then convince management that systems need to be moved or built in the cloud. There are some applications that fit better in the cloud, but not all of them. I certainly don’t want sensitive information in the cloud, at least until we work out some of the legalities for who owns, controls, and responds to subpoena about data.

    The applications that make sense seem to be those that are distributed with lots of paying clients. I ran across a blog that looked at some of the companies that have moved into the Azure cloud with pieces of their businesses and been quite pleased with the results. Quite a few of these are cost-scalable by clients, meaning that a new user or customer is paying some fee that makes it economical to add new databases and servers for that client. If they leave, you can shut down their server.

    It’s definitely easier to start an application in the cloud rather than move an existing one. The architecture is different from an on-premise application, and that means code changes. Not something many companies want to engage in, given the past success of such projects. However the cost savings can be significant for new projects, if those new projects involve investment in equipment, facilities, people, or some combination of all of those.

    However one of the biggest items I see mentioned why companies like cloud platforms like Azure is the speed of deployment. To me that means that the IT infrastructure people, from storage to admin to DBAs, are falling down on the job. There’s no reason we can’t deploy new machines as quickly as Azure these days.

    Steve Jones

    The Voice of the DBA Podcasts

    We publish three versions of the podcast each day for you to enjoy.

  • Helping MS Help Us

    Vote today! See if we can help MS help us.
    Vote today! See if we can help MS help us.

    I think SQL Server is a fantastic product. I’ve been using the platform over 20 years, across 8+ versions and 2 operating systems. It’s provided me with a great career, and one I’ve enjoyed. However the software isn’t perfect, with numerous bugs and holes. The Connect system was introduced years ago as a way for people to submit bugs and suggestions, vote on them, and provide feedback for Microsoft.

    Does it work? I think Connect is broken, and while I sympathize with the volume of suggestions that Microsoft must deal with, I’d hope that they actually listen, consider, and act on something other than documentation bugs. Those seem to be the only ones acted on lately.

    Most of the items submitted are sent in by individuals, and they get one or two votes, but that’s usually from the submitter’s friends. Many of the items are, IMHO, not important enough to submit to Connect. My guess is they cause plenty of people in Microsoft to view Connect as a slop bucket where most users toss every complaint they have about the product. I sometimes wonder if items are submitted by users before they even spend a few minutes on a search engine trying to resolve their problem.

    However there are some great suggestions, and this is one. It’s asking for a new virtual table: Errors. This is in addition to the inserted and deleted tables. It’s not just that I think so, but there are over 300 people that have voted it up. It’s the number 6 highest voted item on Connect. Users see value in this, especially data warehousing users. If you read the comments, this thread, or this blog, you might agree. If so, then I’d encourage you to vote it up. If not, vote it down. Either way, leave a comment on why this would help (or not help) your work. I don’t know if this will change Microsoft’s mind on the issue, but it would be interesting to see if they respond with anything other than there are “resource constraints” preventing this from being implemented.

    While you are at it, look over the list of top voted on items. There are some good ones, like fixing this error, create or replace, and enhancing the OVER clause. If any of them would make your job easier, let Microsoft know. They always say that customer problems and situations give them reasons to enhance the product or fix issues. Let’s give them a chance to actually prove that statement is true. Participate in our experiment and vote today.

    Steve Jones

    The Voice of the DBA Podcasts

    We publish three versions of the podcast each day for you to enjoy.

  • Data Will Drive the World

    Software, and data, will eat the world.
    Software, and data, will eat the world.

    There’s a well known essay from Marc Andreeson that talks about howsoftware is eating the world. There’s a lot of truth to this, in my opinion, and it becomes very important for more and more people to realize that software is going to become more and more of a part of their lives, all parts of their lives. Whether in business, in your personal life, in government or anywhere else, software is going to increasingly be used to interact with the world. This will bring about many opportunities for people in technology to help shape the way those interactions affect all of our lives.

    However it’s not just software that’s important. The data that drives this software is arguably more important than even the software. This data drives the software algorithms to produce some result or action. Hacking the data to change values can even change the results from the software. Different software algorithms might interpret or react to data differently. If data isn’t more important, it’s equally important as the software that processes it.

    We are seeing the additional recording and gathering of data allowing all sorts of new software to be written. Some of the software is performing jobs that humans have do in a faster and perhaps more efficient way. Other software is helping people understand the world in a way they might never have seen it. We even have software replacing other  software jobs. That feels quite surreal to me, but it’s made possible by more and more data being available to help guide the development and growth of systems.

    We are in a good business for the future, with the growing needs to process, manage, and manipulate more and more data in the future. It’s up to us to learn the skills we need to do this efficiently, and in ways that can’t easily be automated.

    Steve Jones

    The Voice of the DBA Podcasts

    We publish three versions of the podcast each day for you to enjoy.

  • Memorial Day 2013

    Graves_at_Arlington_on_Memorial_DayToday is the Memorial Day holiday in the US. I’m sure many of you have the day off, and I hope that you enjoy it, while also remembering those that have given their lives in military service.

    For those of you outside of the US, I have some humorous (I hope) bloopers you can enjoy.

    Steve Jones