Voice of the DBA

Category: Editorial

  • A Patch Disaster

    This isn't the dialog box you want to see on a server.
    This isn’t the dialog box you want to see on a server.

    Can you imagine sending the wrong patch to the wrong machines? That’s what happened with an Australian bank. Am OS patch was sent to many more machines than it should have. The patch was designed for desktops, but managed to get deployed on servers and resulted in some sort of software corruption.

    SQL Server users are fortunate that we rarely have security patches for our platform. There are cumulative updates every other month, but the majority of them aren’t aimed at SQL Server, and many of them may not even be required for the host Windows OS. Your organization’s policy may require the OS patches, and if they do, you should be aware of when and how they are being deployed.

    Even if patches are not supposed to be deployed to your servers, you should plan on being aware of the deployment date. You never know when an administrator will make a mistake and deploy a patch to your database server and necessitate a restore. I would recommend that you double check your backups and ensure restore scripts are handy on any patch day.

    Our computer infrastructure becomes more complex all the time. At the same time, many of us become more specialized, working in a more focused area, and counting on others to manage the parts of our system we do not have time to worry about. People will make mistakes, and we should ensure we can recover our systems from those mistakes.

    Steve Jones

    The Voice of the DBA Podcasts

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

  • Is Big Data Good for Data Professionals?

    Big Data isn't well defined, but it makes a lot of people sit up and pay attention.
    Big Data isn’t well defined, but it makes a lot of people sit up and pay attention.

    There’s a lot of hype around “big data”, a term being thrown around so much in the media that I’m not really sure what it means anymore. Is 1TB “big data”? Is 10,000 transactions/sec big data? Or does it mean that you have more data than your systems can handle, causing queries and reports to run slow?

    I almost hope it’s the latter. I hope that our managers start to think that when our systems run slower that we’re dealing with big data, and we need more resources. The whole big data phenomenon could be a way for data professionals to start a new hardware renaissance, where hardware budgets grow and we begin to replace our current systems with bigger, faster servers.

    Or perhaps it’s a way to offload some of the system administration for individual servers and move to cloud services. I don’t think that’s necessarily a bad move for many DBAs as it would allow them to focus on data management, and information extraction rather than dealing with storage and hardware management.

    Big Data is in the news, and it’s being used by vendors to sell new products and services. From Hadoop to new SANs to BI interfaces, there’s no shortage of places where the term “Big Data” might be used to try and shorten the sales cycle. As a data professional, it’s important that you understand what your needs really are, and if the term is being in a way that actually provides some value to your company for the money spent.

    Steve Jones

    The Voice of the DBA Podcasts

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

  • Statistical Protection

    Statistics are being used more and more, and many of us don't understand the lack of security, even in anonymized data.
    Statistics are being used more and more, and many of us don’t understand the lack of security, even in anonymized data.

    The things people can do with data is amazing. I remember reading about the anonymous data set released by Netflix and how some of the people were identified based on other, related actions on the Internet. This de-anonymization, while scary, was amazing to me. There have been other, related reports of similar “attacks” taken against other data sets. These reports worry me that we will have more and more data security issues in the future, not less.

    I ran across an article that talked about protecting data in statistical databases. These are the databases that contain data from multiple sources, and are used to analyze the information from these sources. The security of these databases becomes important when the data contains information about individuals that we consider sensitive. Interestingly enough, it seems that the security protections being used are query restrictions.

    However these restrictions are the reverse of what we might expect. There might be minimum restrictions on the number of rows returned, to try and prevent information about a specific individual from being returned. There are also limitations on the types of queries that can be run, usually requiring aggregate functions in the query, and restricting which aggregates are allowed.

    This is definitely an area of our industry that needs more work and research. Lots of organizations, especially government organizations are being called on to open their databases up to the public, and many of them are doing so right now, allowing queries of their statistical databases. This might improve the use of this information by the public, but there are plenty of ways in which this data could be potentially misused. If your companies wants to open some of your data to clients or customers, you might raise the concerns with possible abuses of the database and ask that time and effort be included to try and secure the data, possibly by implementing query restrictions.

    Steve Jones

    The Voice of the DBA Podcasts

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

  • The $50,000 Laptop

    Hopefully you won't lose this amount of data.
    Hopefully you won’t lose this amount of data.

    I wrote a long time ago about the value of the data on a laptop being worth more than the hardware. That’s certainly true for me, and I very much worry more about losing the data on my devices than the any of the devices themselves. I use sync services to keep a backup of most things, but I still worry about losing any of my bits.

    There was a large study competed recently, called The Billion Dollar Lost Laptop Program, which examined 329 organizations. The idea was to find out the economic costs of lost laptops related to various public and private entities. The conclusion? The average value of a lost laptop, just one laptop, is $49,000. The conclusions say that least expensive part of losing a laptop of replacing the hardware.

    As we would expect, most losses occurred away from the office, however it’s not known how many of these losses might be targeted thefts. Many of the losses did occur through theft, which is disturbing when most of these disks did not have encryption in place. This was true even when confidential data was contained on the laptop.  It is nice to see that companies that realize they were targets of theft, as opposed to losses, typically do use encryption.

    As the study shows, there are a lot of costs that go into replacing a laptop. Many people don’t think of all of these costs, and even if the costs are double what they should be, these are still substantial costs for companies to absorb. The costs will only go up in the future, especially as more and more people move from desktop workstations to laptops.

    The one positive note? Encrypting the laptop almost cuts the loss in half. A good reason to require encryption on all laptops.

    Steve Jones

    The Voice of the DBA Podcasts

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