Part of the argument is the Uber is building product infrastructure of their own - Amazon built AWS as part of their everyday working, Uber could roll out delivery, autonomous tech or even their financial services as separate strands once they've reached pervasiveness.
When Uber started, they basically reinvented half of Google/Amazon since the cloud was no where near as mature as it was now. They were running their own data centers. It takes a lot of people to rebuild robust NoSql DBs, stream processing, container management, monitoring and the like from scratch.
Could Uber be built with a lot less now days? Absolutely, but it's also an interesting balance of even if they could, if their cloud bill for X service is 10m a year, and they think their own team of 4 could build X internally for less then that...why not do that?
Yep. Always a bit dangerous to go up against AWS and similar. My hope here is that this product is too niche for the major cloud vendors to invest in. But since Uber is building stuff themselves that assumption may be wrong.
A “golden partner” model makes a lot of sense, thanks.
As an interesting side note it was also in the article's last paragraph about Uber not using Public Clouds such as AWS. Is that becoming more common or less?
Is there any other examples of enterprises seeing benefits of running their own services?
Referring to this part:
...they chose to largely manage their own online servers. While that led to some rocky moments early on, people who have worked at the company have said Uber was able to save substantially on costs.
Uber has scale, they are processing $15+ billion plus in rides a year. Millions of customers and millions of drivers that need tools built to service them.
Uber has always struck me as something of an engineer's sandbox. They hire too many engineers who build (and re-build) over-engineered solutions, and who are paid a lot. They have thousands of microservices for example.
Amazon's infrastructure is not easily replicated. As far as I can see Ubers infrastructure is essentially their app and whatever routing/mapping/job matching algorithm they have. Both of which could be replaced by a better implementation relatively easily.
Uber didn't exist before 2009. They got a bunch of people in a room and wrote some code, and now they're successful. Greenfield development at it's best.
Delta started off flying planes, and has been around longer than computers existed. Like many established businesses, they have a mix of new and legacy technologies, and can't just copy everything up to AWS. They'd have to dedicate years of time and effort over many employees to rewrite some of their systems, while still maintaining their existing ones in parallel until it's safe to migrate off. All the while they have an actual business to run, with real revenues, real expenditures, and IT is just (an important, but costly) piece in the big picture.
These companies are neither incompetent, nor malicious. They just have to find the money and time to get done the enhancements they'd like to their systems, and change doesn't happen overnight. Likely, with this awkward generator fire, they'll try to hasten their efforts.
Uber isn't an on demand ride service. They're a logistics company. Also uber isn't waiting for autonomous vehicles to come along and save them, they're one of the industry leaders in developing autonomous systems, investing/acquiring autonomous systems, and collect more driver/vehicle data than any other company in the world. It's not just civilian transportation either. They've already launched an autonomous commercial shipping service. If there is an item, person, or service that could potentially need to be transported from point A to B, it's a safe to assume that Uber has already started collecting logistics data on it and will(or already has) raise funds to begin implementing a network to service said logistical situation. This all may sound a bit extreme at first glance, but look at amazon. Amazon isn't a revolutionary e-commerce market place, they're a revolutionary logistics company. It's the same story as uber just in a different phase. Years of massive losses and outside funding. Both companies fully understand the value/defensibility of heavily investing in operational logistics at massive scale.
Uber has: No CEO. No CFO. No COO. No General Counsel. Recent turnover at heads of business, policy and communications, engineering, and product.
Amazon or Google or any other cloud provider could spin up a service that pairs drivers with ride-seekers. This kind of software is no longer a secret sauce and they have the infrastructure to scale instantly and the customer base to market it easily.
Not much. Uber largely runs its own data centers with some multi cloud. Travis Kalanick
was paranoid that Uber would end up competing with Amazon or Google eventually and did not want to be beholden to them for tech infrastructure.
I believe Lyft is entirely on AWS but they are much smaller overall because they’re only in a limited number of markets.
reply