September 2016 Archives

[Disclaimer: I do not work for Google or Twitter; I have no investments in Google or Twitter. ]

Update: 2016-09-25: Someone pointed out that a better title considering what I'm saying is, "Google should save Twitter as an act of charity".

Google should buy Twitter. (link to rumors here)

Twitter isn't a good "MBA runs the numbers" acquisition. However could be used as a showcase for GCE. It would more than justify itself. In fact, the financial losses might be off-set by the marketing value it provides to GCE.

As part of integrating it into the internal Google stack, they should require their engineers to rebuild it on the Google Cloud Engine platform. GCE scales crazy-good. Twitter has a history of scaling problems. If Google could run it on the Google Cloud Engine, and show that it scales, it would be great advertising.

Google needs GCE to succeed (but that's for another blog post... or you can read Appendix B ofhttps://the-cloud-book.com .. especially the last few paragraphs.)

How difficult would it be to rebuild Twitter on GCE? I think it would be easier than you'd imagine. Every talk I've seen by Twitter engineers at conferences is about technology that (and I don't mean this with any disrespect) is reproducing something in Google's stack. Most of those technologies being re-invented are available in GCE now, and the rest really should be. In fact, if they aren't available in GCE they should be. The project of porting Twitter to GCE would generate a list of high-quality feature requests. Interestingly enough, the re-invented technologies don't seem to be as scalable as Google's original. Oh, and it seems like a lot of people re-implementing those technologies at Twitter are ex-Google employee so ... you have that.

Sadly the few Google executives that I know think that Twitter is a joke, nobody uses it, and isn't worth saving. I disagree. I think it is "the world's chatroom". If you think Twitter "doesn't matter" then why does every news program, TV show, and billboard list a Twitter handle? (Hint: they don't list G+ handles... does G+ even have handles?)

So, in summary:

• Google should buy Twitter.
• It would help save this important resource that the world finds very useful.
• It would be the best showcase of GCE evah.... which is something Google needs more than the revenue of Twitter.
• Sadly Google executives dis Twitter as a niche application that a very small number of people find compelling. (Spoiler alert: I think they're wrong)

I wonder what will happen.

Tom

NOTE: This article was written by Thomas Limoncelli and included no involvement by current or past co-authors or friends.

I'll be giving a new talk, "Teaching DevOps to Ops without Devs (and so can you!)" at Velocity NYC, October 19-22, 2016. Please use this link for information or to register..

Pearson is doing their annual "Back to Business" sale until Monday, September 19. You can save 35-45%, which is a big deal IMHO.

The Practice of Cloud System Administration is 35% off, or 45% off if you buy 2 copies. Buy one for yourself and get a copy for a friend for their birthday. Just use this link to receive the discount.

You can also get a copy of the Cloud Administration book plus the new 3rd Edition of The Practice of System and Network Administration (when it ships in November) and save 45% if you use this link and enter offer code "B2B".

These offers end on Monday. Act now!

(NOTE: I'm not saying our books are great birthday gifts. However think about it this way. What's the worst that could happen? If your friend is offended and never talks to you again at least you'll have more time to read your book. Isn't that what you wanted in the first place?)

15 years ago today (or August 24, 2001 depending on who you talk to) the first edition of The Practice of System and Network Administration reached bookstores.

We had been working on the book for 2+ years, having first met during a Usenix conference in 1999. Writing it was quite an experience, especially since this was before voice-chat on the internet was common, and we were on different continents (Christine in London and Tom in New Jersey). We collaborated via email, used CVS for our source code repository, and we had monthly phone calls (which Tom dialed from work... Thanks, Lucent!). At the time collaborating this way was considered quite radical. Most authors emailed chapters back and forth, and had a hell of a time with merge conflicts. Our publisher was amazed at our ability to collaborate so seamlessly. This kind of collaboration is now commonplace.

The book did quite well. We've sold more than 38,000 copies, in 3 editions (2001, 2007, and 2016), and many printings. It is available on softcover, ebook, and as a web page. It has been translated into Chinese and Russian. In 2005 we received the Usenix SAGE/LISA Outstanding Achievement Award. The 2nd and 3rd editions added an additional co-author, Strata R. Chalup. Strata's experience and project management skills have been a real asset. We've worked with many editors and other production people at Pearson / Addison-Wesley, starting with Karen Gettman who originally recruited us. Thanks Karen, Catherine, Mark, Debra, Kim, Michael, Julie and many others!

We've had a number of surprises along the way. Our favorite was visiting Google (before Tom worked there) and shown a supply closet full of copies of the book. It turned out all new Sysops members were issued a copy. Wow!

Most importantly, we've received a lot of fan mail. Hearing how the book helped people is the biggest joy of all.

In November, the 3rd edition will reach bookstores. We're very excited about the new edition. It has over 300 pages of new material. Dozens of new chapters. It is more modern, better organized, and has a lot of great new stories. You can pre-order the book today. You can read drafts online at SBO. Visit the-sysadmin-book.com for more info. (This is not to be confused with the sequel book, The Practice of Cloud System Administration.)

Thank you to everyone that has purchased a copy of The Practice of System and Network Administration. We really appreciate it!

The last time you joined a new operations/devops/sysadmin team, what make it easy or difficult to get started?

For example... if procedures aren't documented, it can be very difficult to know how to perform them. The rest of the team does them by memory, but you are spending your time asking for help, or guessing your way through them. Well-documented procedures (or even tiny bullet lists) make it easy to be self-sufficient quickly.

What have you found made assimilating into a new team difficult? What have you seen teams do that made it easy? What did you wish a team had done to make it easy?

Please answer in the comments.