Author: way0utwest

  • Sharing Data

    One of the things that’s written into various computer laws is that unauthorized access of a computer system with another’s credentials is a crime. That’s been drilled into me at various organizations where no one was supposed to share any credentials. That is something many sysadmins drill into users, while sharing their own system level credentials for accounts like root, sa, system, and more. I’ve seen no shortage of these accounts in use across multiple services or multiple people from those that would admonish a manager for sharing their password with an assistant.
    This week Troy Hunt wrote about politicians sharing passwords, and the problems with doing so. It’s an interesting read, and certainly points out that the expediency of having users share a workload has plenty of downsides in accountability and auditing of actions. I think there’s little excuse for sharing security credentials in UK gvoernment as there are other solutions to handle this issue. I am more sympathetic in real time environments, like hospitals, where the login process might literally cause a death in the event of a delay.
    Leaving aside the authentication aspect, we often share data among individuals inside of an organization. In fact, outside of sysadmins, there might not be many people that really understand who should have access, let alone who has access, to some data. In fact, over time it seems that most organizations tend to lean towards allowing an ever-growing number of people to access data in file shares. While we might prevent database access and grant/revoke this at times, the output from our systems often ends up in Excel sheets or other files and people that might not have direct access still see the data.
    The real physical world is just as bad, since many people may leave data lying around on desks or tacked to a wall. Just like credentials on a post-it, we have lots of data available for others to read, though physical access is required. However, have you thought about how many people have physical access? It’s not just your co-workers, but also janitorial staff, tradespeople, and others likely wander regularly through your office spaces.
    Security is a tough battle, one that is interesting in that most of the time we don’t need much more than good passwords. Most people don’t have the time or inclination to deal with their own data, much less yours. However, when an attack is targeted on your organization, from outside or within, it’s extremely difficult to ensure data won’t get lost.
    I don’t have a great solution, but I do think that there are good reasons to limit access to data on our systems, not the least of which is auditing and accountability. Beyond that, however, we have to hope that our users have some judgment about with whom they may share reports and other data.
    Steve Jones
  • SSMS Updates in 17.4

    SSMS 17.4 is out, and if you haven’t upgraded from SQL 204, 2012, 2008, etc., you should do so. You can run your bundled SSMS side by side with the new versions, but it seems that the 17x versions are improving andbecoming more stable.

    You can download the new version and use it for free. That alone should be a reason to update your workstations.

    This update includes a few new interesting items.

    • Vulnerability Assessment – Scanning for security and misconfiguration issues
    • Always On Dashboard – New latency reports
    • Showplan – some fixes and new operator icons
    • XE Profiler – Very cool to see this being enhanced.

    There are other updates and fixes as well. I’m sure there are some bugs, but for the most part, I think that SSMS 17.x is the way to go forward.

  • Singular or Plural

    There are all sorts of interesting debates that developers will have about programming. One of those I’ve written about in the past is Spaces v Tabs. There’s a lot of discussion about that one, and certainly no share of strong opinions. I’m sure there are plenty of other development paradigms and habits that will create debate, arguments, and perhaps strain some working relationships.

    While I’m not looking to upset anyone, I ran across a discussion recently and thought this would make an interesting debate. If you’ve got an opinion, please share how and why you might choose to follow your convention. Examples are helpful and may enhance the reasons why you go choose to build systems one way or the other.

    If you examine any sample databases out there, you will run across tables named like this: Customers, Orders, Cities, etc. You might also find Product, Person, Address as well. If you are paying attention, you might notice that my examples are both singular and plural in their form. There are no shortage of debates on the topic, but I’m wondering what many of you think.

    I’ve tended to build tables with plural nouns in the past, but I think that’s because the first few people that taught me did that. As I’ve read more and listened to others explain their design decisions, I’ve embraced singular names. After all, the entity being modeled is often a single instance of a type. A Person or a CreditCard, not a series of them, though we could certainly argue the table is a set of people, so use that.

    Ultimately I’m not sure that it matters much in any particular application. We certainly have databases that using each convention, and probably a few that use both inside the same schema. I think choosing an entity name that is easy to understand is important, and maybe the idea of singular or plural names matters less. After all, as long as you’re not naming tables F42 and H1492, everyone will get used to the convention.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio (3.5 MB) podcast or subscribe to the feed at iTunes and Libsyn.

  • Interviewed

    I got sent an interview by Mohammad Darab, a fellow SQL Server professional in Northern Virginia. I thought the questions were interesting, and spent a bit of time on them. You can read the interview here: https://mohammaddarab.com/interview-steve-jones/