web services

NVIDIA begins shipping the A100, its first Ampere-based data center GPU

NVIDIA announced today that its NVIDIA A100, the first of its GPUs based on its Ampere architecture, is now in full production and has begun shipping to its customers globally. Ampere is a big generational jump in NVIDIA’s GPU architecture design, providing what the company says is the “largest leap in performance to date” across all eight generations of its graphics hardware.

Specifically, the A100 can improve performance on AI training and inference as much as 20x relative to prior NVIDIA data center GPUs, and it offers advantages across just about any kind of GPU-intensive data center workloads, including data analytics, protein modelling and other scientific computing uses, and cloud-based graphics rendering.

The A100 GPU can also be scaled either up or down depending on the needs, meaning that you can use a single unit to handle as many as seven separate tasks with partitioning, and you can combine them to work together as one large, virtual GPU to tackle the toughest training tasks for AI applications. The ‘Multi-instance GPU’ partitioning feature in particular is novel to this generation, and really helps emphasize the ability of the A100 to provide the most value for cost for clients of all sizes, since one could theoretically replace up to seven discrete GPUs in a data center if you’re already finding you have some headroom on your usage needs.

Alongside the production and shipping announcement, NVIDIA is also announcing that a number of customers are already adopting the A100 for use in their supercomputers and data centers, including Microsoft Azure, Amazon Web Services, Google Cloud and just about every significant cloud provider that exists.

NVIDIA also announced the DGX A100 system, which combines eight of the A100 GPUs linked together using NVIDIA’s NVLink. That’s also available immediately directly from NVIDIA, and from its approved resale partners.

Cloud Foundry renews its focus on developer experience as it looks beyond the enterprise

The Cloud Foundry Foundation (CFF) just went through a major leadership change, with executive director Abby Kearns stepping down after five years (and becoming a CTO at Puppet) and the CFF’s CTO Chip Childers stepping into the top leadership role in the organization. For the most part, though, these changes are only accelerating some of the strategic moves the organization already made in the last few years.

If you’re unfamiliar with the open-source Cloud Foundry project, it’s a Platform-as-a-Service that’s in use by the majority of Fortune 500 enterprises. After a lot of technical changes, which essentially involved building out support for containers and adding Kubernetes as an option for container orchestration next to the container tools Cloud Foundry built long before the rise of Google’s open-source tool, the technical underpinnings of the project are now stable. And as Childers has noted before, that now allows the project to refocus its efforts on developer experience.

That, after all, was always the selling point of Cloud Foundry. Developers stick to a few rules and, in return, they can easily push their apps to Cloud Foundry with a single command (“cf push”) and know that it will run, while the enterprises that employ them get the benefits of faster development cycles.

On the flip side, though, actually managing that Cloud Foundry install was never easy, and required either a heavy lift from internal infrastructure teams or the help of outside firms like Pivotal, IBM, SAP, Suse and others to run and manage the platform. That pretty much excluded smaller companies, and especially startups, from using the platform. As Childers noted, some still did use it, but that was never the project’s focus.

Now, with the Kubernetes underpinnings in place, he believes that it will become easier for non-enterprise users to also get started with the platform. And projects like KubeCF and CF for K8s now offers a full Cloud Foundry distribution for Kubernetes, which makes it relatively easy to use the platform on top of modern infrastructure.

To highlight some of these changes, the CFF today unveiled its new tutorial hub that will not just explain what Cloud Foundry is, but also feature tutorials to get started. Some of these will be hosted and written by the Foundation itself, while community members will contribute others.

“Our community has created a learning hub, curated by the Cloud Foundry Foundation, of open-source tutorials for folks to learn Cloud Foundry and related cloud native technologies,” said Childers. “The hub includes an interactive hands-on lab for first-time Cloud Foundry users to experience how easy the platform makes deploying applications to Kubernetes, and is open for the community to contribute.”

Online printing site Doxzoo exposed thousands of customer files

Doxzoo proudly says on its website that your “documents are in safe hands.” But for some time, that wasn’t true.

The U.K. printing company left its customer files on a cloud storage bucket, hosted on Amazon Web Services, without a password. Anyone who knew the easy-to-guess bucket name could access the massive trove of customer files. By the time that the company secured the bucket, it contained more than 250,000 customer-uploaded files.

When reached by email, Paul Bennett, one of the company’s directors, confirmed the exposure.

“The data we store [with Amazon] is solely the files we use for printing their documents and we have a clear privacy policy on our website to cover how this data is held,” said Bennett.

“We frequently review processes and technical architectures to ensure we adhere to current best practices. We are committed to providing the best possible service to our customers and take the security of their personal data very seriously,” he added. “We have already sought guidance from the ICO on our data security and the precautions we take.”

But a spokesperson for the U.K.’s Information Commissioner’s Office (ICO) said it has not received a notification of a security lapse from Doxzoo.

“People have the right to expect that organization’s will handle their personal information securely and responsibly,” the ICO spokesperson said. “Where that doesn’t happen, people can come to the ICO and we will look into the details. When a data incident occurs, we would expect an organization to consider whether it is appropriate to contact the people affected, and to consider whether there are steps that can be taken to protect them from any potential adverse effects.”

Companies that fall foul of European data protection rules can be fined up to 4% of their annual turnover.

At the time of writing, Doxzoo has made no mention on either its blog or its social platforms about the security lapse.

Doxzoo finds itself in similar company to Rallyhood, a Sprint contractor, the Democratic Senatorial Campaign Committee, FormGet, Mixcloud, and Samsung, all of which have in the past year left sensitive data online by mistake.