5 things to read today: From hard drives to hybrid cloud

Sorry this is late. I'll cut to the chase and just serve up the five most interesting things I read t
ARCHITECHT
5 things to read today: From hard drives to hybrid cloud
By ARCHITECHT • Issue #130
Sorry this is late. I’ll cut to the chase and just serve up the five most interesting things I read today.
Google and Pivotal partner on hybrid container clouds: This is essentially a commercial version of the Kubo project Pivotal and Google built earlier this year for running Kubernetes on Cloud Foundry. The core is a new product called Pivotal Container Service, which manages Kubernetes clusters on either VMware or Google Cloud. Yesterday’s VMware-AWS announcement got a lot of coverage, and for good reason, but this Pivotal-Google relationship could provide Team Dell with better long-term prospects. In other news, VMware and Pivotal both joined the Cloud Native Computing Foundation, which GeekWire’s Tom Krazit says is proof it’s becoming the hub of enterprise IT.
Target is the latest retailer to move off AWS: Retail defection/prohibition alone won’t make or break AWS, but it’s a trend worth keeping an eye on as the Amazon platform expands. We’ve already seen Google and Walmart partner on the smart-home front, and you could envision competitors across Amazon’s other businesses looking for other partners, too. If it happens enough, AWS might really have a situation on its hands.
Backblaze releases its latest hard drive analysis: This quarterly report is always a good at what hard drives are working best for storing lots of backup data. If I’m reading this correctly, HGST drives are very reliable over long periods. The bigger picture here, though, is that Backblaze made a name for itself with open-source storage hardware, and its focus on price-performance has made the company profitable even at very low prices. It’s now probably the best/safest bet in town for consumer backup, and has a growing S3-type cloud storage service, as well.
Salesforce built a natural-language query system using deep reinforcement learning: Salesforce has a built an impressive research team, and its willingness to let them publish and open source their work should help it continue to grow. This new system, called Seq2SQL, uses reinforcement learning and SQL under the covers, but presents users with a standard search bar. There are a lot of companies working on natural language queries and raising gobs of money for it, but Salesforce could make a big dent in that market if can do a good job productizing this work. 
Gigster raises $20 million to help enterprises build AI apps: Well, not just AI apps, but that’s obviously going to get a lot of attention in the years to come. More importantly, though, is how the company is trying automate the process of defining a job and securing a team of freelancers to build it. If you buy into the argument that most successful AI systems and applications will be custom-built, then you have to like a company focusing on improving that part of the process.

Sponsor: Bonsai
Sponsor: Bonsai
Artificial intelligence
People seem impressed with the natural translations that DeepL is able to produce. I’m not certain about the market for translation services, but there are several reasons why competing against Google et al seems difficult, even if your models are better. Maybe there’s an acquisition in the works.
This approach seems better than previous ones I’ve seen. And at any rate, there’s something about medical imaging specialists and researchers that’s more reassuring than a startup.
One thing that’s particularly interesting here is the footprints both companies have in Asia. If you want to learn more about Skymind, check out my recent podcast interview with founder and CTO Adam Gibson.
This seems like a fair assessment of the benefits of classic machine learning versus AI for market prediction, as well as the importance of human oversight and the effects of regulation on adopting AI.
This is a huge opportunity for cloud providers like IBM, as the researchers note. They need models that models that can more easily address the varying workloads of varying customers, without doing a custom project every time.
arxiv.org  •  Share
Sponsor: DigitalOcean
Sponsor: DigitalOcean
Cloud and infrastructure
This is a nice analysis of everything VMware announced this week (and previously) and how the company is thinking about hybrid clouds. Somebody suggested to me on Twitter that AWS might stand to make more revenue from its VMware partnership than I suggested yesterday (because TCO of VMware is more than just licenses). It makes some sense, but I still contend that the long-term move will be away from VMs and toward containers, microservices, etc.
This makes a lot of sense because, well, it sounds like very few folks use it—plus a host of other reasons. The post is really good, though, as the authors pretty clearly feel torn about it.
I think this is wishful thinking—the interviewee here is from Dell—and it really depends who ends up building out edge locations (i.e., cloud providers or more traditional buyers). 
This is a short take by a Gartner analyst that makes an important point: technologies don’t need to reach mass adoption to force a change in the way people think about architectures and build products.
There’s a logical argument that containers will eventually give way to functions as the unit of compute, although they’ll still launch as containers. Here’s a project to do just that.
github.com  •  Share
If you haven’t heard of CrateDB, it’s a database designed to ingest and query streaming data in real time. This is a good writeup of what its creators are trying to do.
Researchers have created a new method for producing qubits that are less error-prone. With all the advances coming out of labs, though, you have to wonder what’s actually being implemented at Google, IBM, Rigetti and other commercial quantum computing companies.
This is interesting work, which aims to make Docker more available in cloud environments and less reliant on its daemon. Mesos provides availability improvements, too, although it wasn’t designed with Docker in mind. Docker Swarm might, too.
arxiv.org  •  Share
If you want info on serverless computing—from pros/cons to economics—this is a great place to start. 
arxiv.org  •  Share
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
I normally shy away from talking about AI-for-marketing startups, but Amplero CEO Olly Downs has a pretty good pedigree in the world of data science. So here you are.
This is important stuff for anybody doing business and storing data in Europe. I’ll continue to share good info on it as I come across it.
I’m not certain this is indicative of anything, but it appears that people still search a lot more for “data warehouse” than for anything having to do with cloud-based options. It’s hard to believe anybody is really looking to buy Netezza over Redshift.
This could be a very big deal if it’s widely accepted. Privacy concerns result in lots of valuable data not being shared, but this approach promises to give researchers data on diseases they’re studying, but nothing else.
Their system, called LevelHeaded, essentially combines into in single engine what normally would be done with two. It’s worth taking seriously for data scientists, given the team behind it. 
arxiv.org  •  Share
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.