ArchiTECHt Daily: HPE should buy a container company

The big news out of HPE's earnings call on Thursday—that a major service provider customer singlehand
ArchiTECHt Daily: HPE should buy a container company
By ARCHITECHT • Issue #24
The big news out of HPE’s earnings call on Thursday—that a major service provider customer singlehandedly dinged HPE server revenue by buying less gear than it committed to—is another in a string of bad omens for the company. Normal enterprises are buying fewer boxes and moving to the cloud, while cloud providers are just buying fewer boxes from HPE. 
If HPE can’t beat the cloud providers, it should at least try to make some money off of them. And buying its way into the container space could be a good way to do it. Hear me out …
According to Gartner, and pretty much everybody else, the server market is on a downhill slide and HPE is getting hit particularly hard. And things aren’t going to get better. Even if the great exodus to the cloud somehow slowed down (it won’t), the great migration to microservices is still happening. This is an oversimplification but, generally speaking, forward-thinking companies want fewer machines running more things inside of containers.
Realistically, many companies actually want to do both of these things—move their workloads to the cloud and build applications using containers and microservices. This makes their applications more flexible, more efficient, more scalable and more portable. 
Portability could be a key factor for the large number of enterprises that have not yet moved to the cloud in any significant manner. Containers and container-management platforms provide, in theory, an abstraction layer that can save companies a lot of rebuilding effort should they want to move applications from one provider to another, or potentially deploy applications across clouds.
(For more on the relationship between containers and the cloud, check out the recent ArchiTECHt Show podcast episodes with Sam Ghods of Box and Peter Yared of Sapho.)
Beyond partnerships and investments (including in my former employer, Mesosphere), HPE doesn’t have a play in either of these markets at the moment. But it should. There should be a lot of money to be made helping large companies move to microservices while they also move to the cloud, especially if you can sell them on the promise of portability (or, for the cynics, on the avoidance of lock-in). 
Let the cloud providers worry about selling infrastructure, while you rake in the margins at the management/OS/application layers. And for as long as selling servers is still a business worth being in, having a strong container-management system baked into HPE gear could be a nice value-add. It makes it more of a solution or a platform sale than just a server sale, if you will.
The OpenStack approach to private cloud software might have been a letdown for HPE, but the approach exemplified by Pivotal, Mesosphere, Docker, Kubernetes (pick a company, or the whole project) and their peers is legit, even if it’s still maturing. This is how people want to build applications and think about infrastructure. Buying its way into an entirely new worldview would be a bold move for HPE, but maybe it’s better to risk burning out than to almost certainly fade away.

Source: Fortune
Source: Fortune
Around the web: Artificial intelligence
… and yet here’s a list of the 50 leading it—culled from a list of 100. We’re all excited about AI, but AI-washing is only going to lead to disappointment.  •  Share
But only about five (5) lines at the moment, so don’t go too crazy. Still, as a tool to augment software developers it’s pretty promising.
If it works, maybe we can all go back to reading comments for the occasional diamond in the rough. What I like is that the system, called Perspective, lets users train models rather than assuming it knows what’s toxic and what’s not.  •  Share
There are some good nuggets in this feature on Facebook’s applied AI team, including around the fake news controversy. But it’s the extra-Facebook applications that are really exciting. 
Hopefully, people are paying attention to research on how hackers could trick machine learning models, and paying attention. Also, there’s some good discussion about the black box nature of AI in here.
Maybe I’m just reading too much GeekWire, but I’m starting to buy into this whole “Seattle is the new Silicon Valley” argument. UW does have some serious computer science chops, though.
Just what the title says. If you’re thinking about using any of these, this is worth reading.  •  Share
On the one hand, this is an interesting, scalable approach to obtaining demographic information using a novel data source. On the other hand, it raises some pretty clear questions around discrimination and other issues.  •  Share
Source: Facebook
Source: Facebook
Source: Facebook
Source: Facebook
Around the web: Cloud and infrastructure
I’m not sure there’s a lot of blame to place here, but I’ll also resist the urge to give Cloudflare kudos for being up front about it. This much shorter explanation, from the BBC, works in “bigly.”
I like this idea a lot. It’s an automated tool for highlighting any licensing issues that might reside in your code. Could be very useful for big companies as they embrace open source.
Yes, this is written by a Kubernetes-based startup exec, but it’s still a pretty good explanation of why companies should look at what’s now a red-hot project. And, really, why they should look at microservices at all.
If you remember the big MarkLogic fiasco, and overall fiasco, around the rollout a few years ago, then this interview has a lot more context.
Want to sponsor this newsletter, or the ArchiTECHt Show podcast? Email for more information.
Around the web: All things data
Called Prophet, it’s a package for R and Python that can automatically generate accurate time-series forecasts, but allows for a good amount of customization. Facebook emphasizes the trustworthiness of the reports Prophet generates.
The partnership is probably not a big deal, but this a reminder that there are a lot of OSS projects and companies out there. For example, Apache Kylin is a SQL on Hadoop project, and Kyligence is the Chinese startup commercializing it.
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.