Archives

Amazon Web Services

Docker partners with AWS to improve container workflows

Docker and AWS today announced a new collaboration that introduces a deep integration between Docker’s Compose and Desktop developer tools and AWS’s Elastic Container Service (ECS) and ECS on AWS Fargate. Previously, the two companies note, the workflow to take Compose files and run them on ECS was often challenging for developers. Now, the two companies simplified this process to make switching between running containers locally and on ECS far easier .

docker/AWS architecture overview“With a large number of containers being built using Docker, we’re very excited to work with Docker to simplify the developer’s experience of building and deploying containerized applications to AWS,” said Deepak Singh, the VP for Compute Services at AWS. “Now customers can easily deploy their containerized applications from their local Docker environment straight to Amazon ECS. This accelerated path to modern application development and deployment allows customers to focus more effort on the unique value of their applications, and less time on figuring out how to deploy to the cloud.”

In a bit of a surprise move, Docker last year sold off its enterprise business to Mirantis to solely focus on cloud-native developer experiences.

“In November, we separated the enterprise business, which was very much focused on operations, CXOs and a direct sales model, and we sold that business to Mirantis,” Docker CEO Scott Johnston told TechCrunch’s Ron Miller earlier this year. “At that point, we decided to focus the remaining business back on developers, which was really Docker’s purpose back in 2013 and 2014.”

Today’s move is an example of this new focus, given that the workflow issues this partnership addresses had been around for quite a while already.

It’s worth noting that Docker also recently engaged in a strategic partnership with Microsoft to integrate the Docker developer experience with Azure’s Container Instances.

Why AWS built a no-code tool

AWS today launched Amazon Honeycode, a no-code environment built around a spreadsheet-like interface that is a bit of a detour for Amazon’s cloud service. Typically, after all, AWS is all about giving developers all of the tools to build their applications — but they then have to put all of the pieces together. Honeycode, on the other hand, is meant to appeal to non-coders who want to build basic line-of-business applications. If you know how to work a spreadsheet and want to turn that into an app, Honeycode is all you need.

To understand AWS’s motivation behind the service, I talked to AWS VP Larry Augustin and Meera Vaidyanathan, a general manager at AWS.

“For us, it was about extending the power of AWS to more and more users across our customers,” explained Augustin. “We consistently hear from customers that there are problems they want to solve, they would love to have their IT teams or other teams — even outsourced help — build applications to solve some of those problems. But there’s just more demand for some kind of custom application than there are available developers to solve it.”

Image Credits: Amazon

In that respect then, the motivation behind Honeycode isn’t all that different from what Microsoft is doing with its PowerApps low-code tool. That, too, after all, opens up the Azure platform to users who aren’t necessarily full-time developers. AWS is taking a slightly different approach here, though, but emphasizing the no-code part of Honeycode.

“Our goal with honey code was to enable the people in the line of business, the business analysts, project managers, program managers who are right there in the midst, to easily create a custom application that can solve some of the problems for them without the need to write any code,” said Augustin. “And that was a key piece. There’s no coding required. And we chose to do that by giving them a spreadsheet-like interface that we felt many people would be familiar with as a good starting point.”

A lot of low-code/no-code tools also allow developers to then “escape the code,” as Augstin called it, but that’s not the intent here and there’s no real mechanism for exporting code from Honeycode and take it elsewhere, for example. “One of the tenets we thought about as we were building Honeycode was, gee, if there are things that people want to do and we would want to answer that by letting them escape the code — we kept coming back and trying to answer the question, ‘Well, okay, how can we enable that without forcing them to escape the code?’ So we really tried to force ourselves into the mindset of wanting to give people a great deal of power without escaping to code,” he noted.

Image Credits: Amazon

There are, however, APIs that would allow experienced developers to pull in data from elsewhere. Augustin and Vaidyanathan expect that companies may do this for their users on tthe platform or that AWS partners may create these integrations, too.

Even with these limitations, though, the team argues that you can build some pretty complex applications.

“We’ve been talking to lots of people internally at Amazon who have been building different apps and even within our team and I can honestly say that we haven’t yet come across something that is impossible,” Vaidyanathan said. “I think the level of complexity really depends on how expert of a builder you are. You can get very complicated with the expressions [in the spreadsheet] that you write to display data in a specific way in the app. And I’ve seen people write — and I’m not making this up — 30-line expressions that are just nested and nested and nested. So I really think that it depends on the skills of the builder and I’ve also noticed that once people start building on Honeycode — myself included — I start with something simple and then I get ambitious and I want to add this layer to it — and I want to do this. That’s really how I’ve seen the journey of builders progress. You start with something that’s maybe just one table and a couple of screens, and very quickly, before you know, it’s a far more robust app that continues to evolve with your needs.”

Another feature that sets Honeycode apart is that a spreadsheet sits at the center of its user interface. In that respect, the service may seem a bit like Airtable, but I don’t think that comparison holds up, given that both then take these spreadsheets into very different directions. I’ve also seen it compared to Retool, which may be a better comparison, but Retool is going after a more advanced developer and doesn’t hide the code. There is a reason, though, why these services were built around them and that is simply that everybody is familiar with how to use them.

“People have been using spreadsheets for decades,” noted Augustin. “They’re very familiar. And you can write some very complicated, deep, very powerful expressions and build some very powerful spreadsheets. You can do the same with Honeycode. We felt people were familiar enough with that metaphor that we could give them that full power along with the ability to turn that into an app.”

The team itself used the service to manage the launch of Honeycode, Vaidyanathan stressed — and to vote on the name for the product (though Vaidyanathan and Augustin wouldn’t say which other names they considered.

“I think we have really, in some ways, a revolutionary product in terms of bringing the power of AWS and putting it in the hands of people who are not coders,” said Augustin.

UK’s COVID-19 health data contracts with Google and Palantir finally emerge

Contracts for a number of coronavirus data deals that the UK government inked in haste with US tech giants, including Google and Palantir, plus a UK-based AI firm called Faculty, have been published today by openDemocracy and law firm Foxglove — which had threatened legal action for withholding the information.

Concerns had been raised about what is an unprecedented transfer of health data on millions of UK citizens to private tech companies, including those with a commercial interest in acquiring data to train and build AI models. While Freedom of Information requests for the contracts had been deferred up to now.

In a blog post today, openDemocracy and Foxglove, write that the data store contracts show tech companies were “originally granted intellectual property rights (including the creation of databases), and were allowed to train their models and profit off their unprecedented access to NHS data”.

“Government lawyers have now claimed that a subsequent (undisclosed) amendment to the contract with Faculty has cured this problem, however they have not released the further contract. openDemocracy and Foxglove are demanding its immediate release,” they add.

They also say the contracts show that the terms of at least one of the deals — with Faculty — were changed “after initial demands for transparency under the Freedom of Information Act”.

They have published PDFs of the original contracts for Faculty, Google, Microsoft and Palantir. Amazon Web Services was also contracted by the NHS to provide cloud hosting services for the data store.

An excerpt from the Faculty contract regarding IP rights

Back in March, as concern about the looming impact of COVID-19 on the UK’s National Health Service (NHS) took hold, the government revealed plans for the health service to work with the aforementioned tech companies to develop a “data platform” — to help coordinate its response, touting the “power” of “secure, reliable and timely data” to inform “effective” pandemic decisions.

However the government’s lack of transparency around such massive health data deals with commercial tech giants — including the controversial firm Palantir, which has a track record of working with intelligence and law enforcement agencies to track individuals, such as supplying tech to ICE to aid deportations — raises major flags.

As does the ongoing failure by the government to publish the amended contracts — with the claimed tightened IP clauses.

The Google contract — to provide “technical, advisory and other support” to NHSX to tackle COVID-19 — is dated March 1, and specifies that services will be provided by Google to the NHS for zero charge. 

The Palantir contract, for provision of its Foundry data management platform services, is dated as beginning March 12 and expiring June 11 — with the company charging a mere £1 ($1.27) for services provided.

While the Faculty contract — providing “strategic support to the NHSX AI Lab” — has a value in excess of £1M (including VAT), and an earlier commencement date (February 3), with an expiry date of August 3.

The government announced its plan to launch an AI Lab within NHSX, the digital transformation branch of the health service, just under a year ago — saying then that it would plough in £250M to apply AI to healthcare related challenges, and touting the potential for “earlier cancer detection, discovering new treatments and relieving the workload on our NHS workforce”.

The lab had been slated to start spending on AI in 2021. Yet the Faculty contract, in which the AI firm is providing “strategic support to the NHSX AI Lab”, and described as an “AI Lab Strategic Partner”, suggests the pandemic nudged the government to accelerate its plan.

We’ve reached out to the Department of Health with questions.

Last month, NHS England and NHS Improvement responded to an FOI request that TechCrunch filed in early April asking for the contracts — but only to say a response was delayed, already around a month after our original request. (The normal response time for UK FOIs is within 20 working days, although the law allows for “a reasonable extension of time to consider the public interest test”.)

Earlier this month, the Telegraph reported that Google-owned DeepMind co-founder, Mustafa Suleyman — who has since moved over to work for Google in a policy role — was temporarily taken on by the NHS in March, in a pro bono advisory capacity that reportedly included discussing how to collect patient data.

An NHSX spokesperson told Digital Health that Suleyman had “volunteered his time and expertise for free to help the NHS during the greatest public health threat in a century”, and denied there had been any conflict of interest.

The latter refers to the fact that when Suleyman was still leading DeepMind the company inked a number of data sharing agreements with NHS Trusts — gaining access to patient health data as part of an app development project. One of these contracts, with the Royal Free NHS Trust, was subsequently found to have breached UK data protection law. Regulators said patients could not have “reasonably expected” their information to be shared for this purpose. The Trust was also reprimanded over a lack of transparency.

Google has since taken over DeepMind’s health division and taken on most of the contracts it had inked with the NHS — despite Suleyman’s prior insistence that NHS patient data would not be shared with Google.