3 things to read today: Facebook's new data store; engineering at Airbnb & Pinterest; and an AWS spinoff?

This week has been—and, well, most every week is—rife with news good, bad, and unclear for Amazon and
ARCHITECHT
3 things to read today: Facebook's new data store; engineering at Airbnb & Pinterest; and an AWS spinoff?
By ARCHITECHT • Issue #132
This week has been—and, well, most every week is—rife with news good, bad, and unclear for Amazon and Amazon Web Services. This is has some folks wondering if Amazon wouldn’t be better off spinning out AWS into its own company, so as to spare it from competitive concerns arising from Amazon’s growing footprint:
Both of these posts are filled with reasons for and against such a move, and Tom Krazit at GeekWire notes that the investor community has been pushing for this for years. Frankly, I think it would be very difficult to do for a variety of reasons, including the effect of Amazon’s business culture on AWS and the profit boost that AWS brings to Amazon. But I suspect we’re only going to hear more about this as customers revolt, competitors ramp up, and analysts and reporters search for reasons to question AWS’s cloud computing dominance.
Here are two other really interesting items from Thursday:
  • LogDevice: a distributed data store for logs (Facebook): Facebook actually released a trio of fairly interesting engineering posts today, but this one takes the take from an infrastructure perspective. The post doesn’t come right out and say it, but I believe LogDevice is another case of Facebook outgrowing Hadoop and its ecosystem—similar to the Beringei time-series database it detailed in February. For good measure, here’s a semi-cynical take on LogDevice for the folks who might actually consider adopting this when it’s open source.
  • a16z Podcast: Engineering intent (a16z): This podcast featuring Airbnb VP of engineering Mike Curtis and Pinterest head of engineering Li Fan is really good. There’s some practical advice in terms of how to think about engineering teams and decision-making, as well as how to capture offline data as a source of true user experience, and how to take advantage of AI.

Sponsor: Bonsai
Sponsor: Bonsai
Listen to the latest ARCHITECHT Show podcast
In this episode, Derrick Harris speaks with Brian Ballard, co-founder and CEO of Upskill, about the company’s mission to make customers more productive via augmented reality. Ballard discusses why the original Google Glass disappeared but never really went away, and why Glass Enterprise is poised to succeed where the consumer version failed. He also talks about the broad ecosystem of Glass competitors, applications and users, and why the network is a critical component of industrial AR architectures.
Artificial intelligence
This is a good look at an under-appreciated benefit of advances in AI and deep learning to companies like Google and Microsoft. So much attention gets paid to things like mastering video games and board games that we sometimes forget how these companies make their money.
I recall Baidu being an early committer to the AMD Epyc GPU, but I hadn’t given much thought to AMD’s true potential to challenge AMD. Here’s an argument why it might.
This is a breakdown of industries where the venture arms of large companies are investing in AI. Among the specific industries or use cases, cybersecurity is tops, followed by IoT.
Sometimes, all it takes is an idea. Here are handful of ways that folks are trying to benefit from machine learning in fields from education to entertainment.
This is a fair argument. I’m not sure it’s a “gaping abyss” as the author suggests, but phones do help add a lot of contextual data.
This is a feature I never have to worry about using on Facebook, but it’s good to know how the company built it. What I wonder is if there’s any opportunity for any startup to challenge Facebook, Google et al on photos anymore.
Language understanding is still a challenging problem, but Google has a new approach, called Transformer, that utilizes some different techniques as well as advances in hardware design. That latter point could matter in the cloud business if better performance is increasingly tied to certain hardware.
If you’ve been following this case and want even more details, here’s a feature about new documents suggesting DeepMind wasn’t as innocent as it let on. I think this situation should be a learning opportunity for anybody wanting access to personal records, but there’s not a lot of good that comes from raking DeepMind over the coals.
The partnership between NASA, Google and D-Wave around quantum computing forges ahead. Meanwhile, Google, IBM and others are at work on alternative quantum models, as well.
Intel is not a dummy, so you know it’s not going to let Nvidia, AMD or anybody else conquer AI without a fight. Making TensorFlow run fast on CPUs people already own is a smart move.
Another good use of deep learning to solve tricky problems that used to require relying on people and paperwork. A great use of satellite imagery, too.
arxiv.org  •  Share
Sponsor: DigitalOcean
Sponsor: DigitalOcean
Cloud and infrastructure
This is a pretty fascinating story and a seemingly great technology for helping organizations identify weak spots in their network security that are spread across the internet.
This is a pretty scathing assessment of Apple’s deal with Iowa, which sees the company getting about $200 million in tax breaks and not obligated to give back much. But from factories to data centers, states seem giddy to lure big names at all costs.
If this was announced during the event this week, I missed it. But here are some details. Again, however, I wonder how many companies will choose to opt for VMware as a platform for containerized workloads rather than choosing something container-native.
Goodbye, DRAM-based caches. Hello, BlueCache, which uses flash memory! I could envision companies like Facebook and others being very excited by this, but what about cloud providers …
This is on my list of podcasts to listen to this week. One of the trio of interviewees about the Kubernetes technical roadmap is Eric Brewer, who’s remarkably adept at bridging technological concerns with business ones. See here, for example.
From Joe Beda, who helped build Kubernetes and co-founded Heptio. By old, I mean it’s from November 2016.
www.cncf.io  •  Share
Possibly naive question: If the federal government is working on data center consolidation, why is northern Virginia such a hot data center market? Perhaps because cloud providers and others are setting up shop to serve SaaS and IaaS to agencies from there?
If you enjoy the newsletter, please help spread the word via Twitter, or however else you see fit.
If you’re interested in sponsoring the newsletter and/or the ARCHITECHT Show podcast, please drop me a line.
Use Feedly? Get ARCHITECHT via RSS here:
All things data
IBM’s acquisition of the Weather Channel was interesting and all, but I never though weather data could be the linchpin of so many partnerships. Now, Salesforce users can pull in weather data from IBM, and IBM Cloud users can connect directly to Salesforce data.
By controlling everything from the sensors to the data stores, Oracle would seem to have a pretty good story for industrial IoT customers. On the other hand, will companies line up to sell themselves to Oracle for the next couple decades, as well?
This is a really interesting take on the upcoming EU regulations, even if it is a bit biased (PageFair is an ad company, and Google and Facebook eat everyone’s ad revenues). It seems like companies should be taking this very seriously, but maybe I’m just having a Y2K moment.
New ARCHITECHT Show every Thursday; new AI & Robot Show every Friday!
New ARCHITECHT Show every Thursday; new AI & Robot Show every Friday!
Did you enjoy this issue?
ARCHITECHT
ARCHITECHT delivers the most interesting news and information about the business impacts of cloud computing, artificial intelligence, and other trends reshaping enterprise IT. Curated by Derrick Harris. Check out the Architecht site at https://architecht.io
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.