ArchiTECHt Daily: Google's flexes its engineering muscle in cloud battle

On any other day, the cloud press might have been going crazy over AWS's new WebEx-style communicatio
ArchiTECHt Daily: Google's flexes its engineering muscle in cloud battle
By ARCHITECHT • Issue #18
On any other day, the cloud press might have been going crazy over AWS’s new WebEx-style communication service Chime, and perhaps AWS’s new Elastic Block Store volumes, as well. But yesterday was not any other day: It was the day that Google announced its long-awaited Spanner database as a service (more detail on that later).
What we’re seeing in the increasingly heated battle between the two companies, encapsulated by yesterday’s news, is that AWS has inherited the DNA of its parent company, betting on retail-style bulk as a winning strategy. Google, meanwhile, is banking on its DNA in distributed systems engineering and betting users will buy into its technocracy.
Jay Kreps—who helped create Apache Kafka (among other things) while at LinkedIn, and current co-founder of big data startup Confluent—summed it up in similar, but slightly different, terms on Twitter:
And yesterday, if my networks and news feeds are any indication, technocracy won. Google trotted out its VP of infrastructure, and CAP Theorem creator, Eric Brewer, to blog about about how Spanner essentially turns his theorem on its head (it also could have chosen Spanner creator Jeff Dean for the job). Google solicited a detailed trial case study by early-access user Quizlet, documenting Spanner performance over 30 nodes. And people ate it up.
Part of this is because Spanner is a remarkable technology, and people are rightly excited about the new Cloud Spanner offering. But I also assume that people—and I mean everybody from journalists to cloud users to CIOs—are just suffering from AWS fatigue, and they’re rooting for anybody who can make things interesting. (Microsoft Azure might be the presumptive No. 2. in terms of cloud revenue, but Microsoft seems content to do its own thing without making much of a splash.)
Don’t get me wrong: Amazon Chime may yet prove to be the best business decision announced yesterday (the cloud is a platform play, after all, and market leader AWS was lacking a conferencing feature) but people really do like what Google is up to. The big question now is whether enough of them favor technical superiority over breadth of products and proven operational efficiency, and by how much.

Source: Google
Source: Google
Around the web: The Cloud Spanner edition
My apologies if you read the intro to this newsletter and were like, “What the heck is Cloud Spanner?” 
In a nutshell, Spanner is a distributed relational database system, developed by Google, that can be distributed across the globe and still provide ACID transactions. From an engineering perspective, it’s a pretty big deal. Cloud Spanner is Google’s new hosted cloud version of it. 
Here is a collection of links from Tuesday that explain in more detail what it is and why people are so excited about it:
Source: Quizlet
Source: Quizlet
Around the web: Security
Glad to see Microsoft leading the charge on trying to prevent cyberwar wherever possible. But you have to wonder where companies will draw the line with respect to their neutrality. 
It’s Ben Horowitz speaking with some smart folks, so you know it’s insightful. Security still not solved, though.  •  Share
Doesn’t seem like such a bad idea considering recent events. I like the inclusion of Apache NiFi, too. However, when I think big data I don’t think HPE.
How much of this is true or how useful it is I do not know. But it sounds pretty badass.
Around the web: Cloud and infrastructure
If it wants to keep customers as they and their applications mature, this is a must-have feature.
Cheaper is better for bare compute resources—and these have better specs than comparable ones from AWS and DigitalOcean—but I bet neither company is losing much sleep.
Last month it was Uber, yesterday it was Pinterest. Not as sexy as distributed databases, but quite important in the age of microservices.
This seems like sound advice, especially if you’re just starting down the path of NoSQL. 
Every foundation becomes a lot less interesting (although arguably more legitimate) when legacy and unknown companies start joining. CapitalOne and Ticketmaster are good additions, though.  •  Share
Historically, Cycle Computing customers used AWS for high-performance computing, so this is a minor win for Microsoft. One workload at a time …
Source: University of Pittsburgh
Source: University of Pittsburgh
Around the web: Artificial intelligence
Advances in AI-optimized hardware have been a recurring in this newsletter, and here is another bunch of examples. The opportunity here is YUGE.
More hardware innovation, from the University of Pittsburgh, this time attempting to increase the size of models that brain-based chips can handle.  •  Share
It uses probabilistic programming, and is marketing itself as (my words, not Gamalon’s) AI for the little guy. For how the company fits into the great AI landscape, read this article from WIRED.
In a way, they’re giving robots common sense about how to get from Point A to Point B, which is harder than it sounds.  •  Share
Around the web: All things data
For some reason, I find it a good sign when the CEO of a large-ish company can discuss why and how it uses certain technologies, which is what Chad Dickerson does here with Hadoop.
Spoiler alert: This is not the funniest thing you’ll read today. But after spending some time working on marketecture, I have an appreciation for the “data cake.” 
Did you enjoy this issue?
The most interesting news, analysis, blog posts and research in cloud computing, artificial intelligence and software engineering. Delivered daily to your inbox. Curated by Derrick Harris. Check out the Architecht site at
Carefully curated by ARCHITECHT with Revue. If you were forwarded this newsletter and you like it, you can subscribe here. If you don't want these updates anymore, please unsubscribe here.