How LinkedIn's Project Inversion Saved the Company

pacopico writes "Shortly after its 2011 IPO, LinkedIn's infrastructure almost collapsed. The company had been running on decade's old technology and needed a major overhaul to keep up with other social sites. As Businessweek reports, LinkedIn initiated Project Inversion to fix its issues and has since evolved into one of the poster children for continuous development and creating open source infrastructure tools. But the story also notes that LinkedIn's technology revival has come with some costs, including constant changes that have bothered some users."

Comments

Dubious story, dubious subject...

[Frosty Piss]

I smell a Slashvertisment... Seriously, LinkedIn? Biggest spammer in my Inbox. Of dubious professional value. Facebook, *please* buy them?

Re:Dubious story, dubious subject...

[i kan reed]

Several points:

1. LinkedIn is actually a sort-of competitor of slashdot's owner. That's a bit of a weird slashvertisement choice(not impossible, but weird).
2. The FTC would come down on facebook like a ton of bricks if they tried to buy out one of the largest other social networks.
3. Remember to report spam on those emails so that someday we might collectively not get them.

Re:Dubious story, dubious subject...

[alostpacket]

Well, there are a few "interesting" gems there. Though it's mostly business fluff. From TFA:

Such companies as Facebook (FB) and Google also have special teams that review the lines of code written by developers. It’s these people who get to decide when a new feature is ready to make its way to their websites. Not LinkedIn. It has one, huge stash of code that everyone works on, and algorithms do the code reviewing. “Humans have largely been removed from the process,” Scott says. “Humans slow you down.”

Uh, Okay. Automated code review? Um, where to begin? I think there an obvious misunderstanding on the part of the author of the article. Surely Google, FB, et. al., do CI and all sorts of automated testing. They just *also* use humans.

Incidentally, Google clearly has more products, thus more specialties and codebases. FB also, to a lesser extent. I dont think the Google Search team is the same as the Google Maps team or the Android team.

LinkedIn is a website, they have an API, messaging, maybe some mobile apps? It's not trivial, but it's probably not very close to the technical complexity of FB, and no where near the technical complexity of Google.

LinkedIn initiated Project Inversion to fix its issues and has since evolved into one of the poster children for continuous development

...by stopping all continuous dev so they could rebuild from scratch...

I think TFA misses the point in a very "PHB way" sadly. They took the time to make the devs happy and give ownership of features to devs. The result was the devs created an environment that was productive and could be continuously updated with less fuss.

To me, this is the poster child for creating a dev focused culture, and taking the time to do things the right way. Which, sadly, is the exact opposite of the conclusion of TFA and the LinkedIn PHB.

Re:Dubious story, dubious subject...

[elloGov]

I drank from the fire-hose and voted this story up.

I work for a dot com older than LinkedIn which is crippled by a similar (worse) monolith legacy webapp. Innovation, efficiency, development cycles, new features and site/product-wide roll-outs are a pain or in some cases impossible. This story gives one perspective/solution to the problem.

Software is a fast moving space, you snooze you lose. As the web matures, I see many more once-prolific trendsetting companies slip into bureaucratic process-driven monoliths milking every bit of value the antiquated software still holds. The wise companies invest in technology and reap the benefits of the initially intangible results of a flexible, maintainable, truly agile technological stack, however, most companies eventually fall into the cycle of:

1. Start-up and innovate
2. Grow and profit
3. Implement n-layers of bureaucratic oversight and process to protect the value
4. stop their ongoing evaluation
5. Loose market-share to newcomers and mavericks in your segment
6. paralyzed with more market-share loss, copy the competition wherever you can and desperately hold onto your scraps
7. keep losing market-share
8. disappear into the abyss

Re:Dubious story, dubious subject...

[tattood]

One is about your ego and broadcasting to an audience of other people who are all busy worrying about their ego and broadcasting to their audience

I'd say that both sites fit that description. LinkedIn has turned into the "Facebook for professionals". It seems like most people's goal on LinkedIn is to connect with as many people as they can to get their "network" as large as they can. I have gotten LinkedIn requests from people that I met at a conference several years ago, and talked to for 5 minutes.

LinkedIn used to be about creating a network of trusted colleagues, that you want to keep in touch with, so that you could get trusted introductions to people you didn't know. If I trust person A, and person A trusts person B, B trusts C; therefore if person C is looking for a job opportunity, then they have a good chance of being a reasonably good candidate. That whole concept seems to have gotten lost in the last few years, and now it is all about having as many connections as you can.

Re:Dubious story, dubious subject...

[ygtai]

I have even gotten (plenty of) LinkedIn requests from people that I have never met and never talked to. Another thing that bothers me (a lot) is from people who know absolutely nothing about my professional capability endorsing my professional capability...

decade's, eh?

[Anonymous Coward]

First of all, it's decades without an apostrophe you doof. Secondly how can a company that's only a decade old run on "decades" old hardware? They bought ten year old computers in 2003?

Re:decade's, eh?

[sribe]

First of all, it's decades without an apostrophe you doof.

Apparently you are unaware that in modern usage an apostrophe no longer indicates possession or a contraction. It now indicates OMG WATCH OUT THERE IS AN "S" COMING UP NEXT!!!"

Cool, no details

[stewsters]

So, an article with no technical details? Cool. What are they doing thats so new?

A while ago I noticed their name on the bottom of this : http://www.playframework.com/

UNImprovements since then?

[whoever57]

That would be about the time that LinkedIn started making the search features LESS effective. For example, in the past, I could review lists of new LinkedIn members that worked for the same companies as I did, at the times that I was there, When I had determined that I did not know them, it would not show me those names again.

The classmates search is completely useless to me. I can no loger add search terms to the search to narrow down the results (I used to be able to do this). All I can do is get the same list of classmates that I have seen before. Since I left university decades ago, I don't have many existing connections to classmates, so a graph search for related classmates is little use to me. I want to search by looking for common courses or interests at the time I was there. Probably, for people only a few years out of college (the Facebook generation), this isn't a problem, since the connections were established while at college.

So, perhaps the infrastructure is better, but from this user's perspective, the site has got worse.

Details Please

[Ghjnut]

Where is all the tech stuff? I want to know what systems were swapped out, what was used in place or what was swapped, what the steps were (did they set up unit tests first followed by architecture changes and scalability testing), what new coding practices they employed etcetera. I'll sum up this horn-tootin session: "LinkedIn had to change to grow, and they did".