Google is working on a new image format called Pik, as evidenced by a collection of newly-uploaded files on the company’s official GitHub profile. The format, according to the project’s readme file, is a “new lossy image format for the internet.” That’s about all we know about it for the time being.
Among the early benefits of artificial intelligence systems, machine learning and natural language processing have the potential to transform customer service processes that are too often frustrating for customers and expensive for web hosts and service providers. A number of tools are currently available to enable companies to overhaul their interactions with customers contacting them for help by telephone or over the internet.
IBM launched what it calls “the next generation of the world’s most powerful transaction system,” IBM Z, on Monday. The technology is able to run more than 12 billion encrypted transactions per day and introduces an encryption engine which it says makes it possible for the first time to pervasively encrypt data from any application, cloud service, or database.
Microsoft last week unveiled plans to bring Windows Server into its semi-annual update release cycle, starting this fall, alongside a new Nano Server image configuration with a 50 percent reduced footprint. The company is stripping the infrastructure roles from Nano Server to optimize the headless configuration option for deployment of container-based environments. While Microsoft revealed the Nano Server changes during last month’s Build conference in Seattle, along with other features coming to Windows Server including Linux container support and plans to offer the server OS into its new semi-annual cycle, the company on Thursday made its plans official and provided some additional details.
Only those opting for the newly minted semi-annual channel can implement any of the new technical features planned for Windows Server, which include the pending stripping down of Nano Server. In addition, organizations will be required to have Software Assurance coverage to access the semi-annual channel releases. The revamp of Nano Server is noteworthy because leading up to its release last year, Microsoft had touted the minimal-footprint deployment option for Windows Server 2016 for its suitability for large clusters in Web-scale application and data center environments. However, Microsoft has since found that the “vast majority” of Nano Server deployments from a workload perspective are running container-based applications based on Docker, Kubernetes, and others. Since container-based workloads do not require the infrastructure components, Microsoft determined that removing them would result in a more efficient server environment and advance the move toward containers.
“Nano Server will be optimized as a container runtime image and we will deprecate the infrastructure roles,” said Chris Van Wesep, Microsoft’s director of enterprise cloud product marketing, “so for anybody who had wanted to do smaller footprint compute and storage clusters, Server Core will be the right implementation to do that.” By deprecating the infrastructure features in the Nano Server option, the removal of that code will make way for Microsoft’s new .NET Core 2.0, “which enables customers to use more of their code in more places [and] make Nano Server the best option for new container-based development,” said Erin Chapple, general manager for Windows Server, in a blog post announcing the new release options.
Microsoft is recommending Server Core for hosting virtual machines as well as containers, which Chapple said can run a Nano Server or Linux container images. The Windows Server update this fall will support Linux workloads via extended Hyper-V isolation, which will allow Linux containers to run without having to deploy two separate container infrastructures to run both Linux and Windows-based applications. As previously announced, Microsoft is also bringing the Windows Subsystem for Linux, (a.k.a. Windows Bash component), allowing application administrators and developers to use common scripts and container images for both Linux and Windows Server container hosts, according to Chapple.
Collectively, these technical changes to Windows Server and the continuous release cycle option associated with it are part of Microsoft’s strategy to bring more consistency to the server OS and Azure. The changes also promote the development of modern cloud apps and migration of legacy apps and systems to these environments using container images. “Many customers don’t realize that Server Core is the base image that runs Azure and Azure Stack,” Chapple noted. “This means the investments we make in Windows Server for Azure can be made available to customers to use in their own data centers. This makes Server Core a great choice for Azure customers who like the idea of consistent technologies between their datacenter and the cloud. One example of this in the upcoming feature update is the cluster sets functionality for increased scale of hyper-converged deployments. We’re also continuing to add security investments such as the ability to quickly encrypt network segments on their software-defined networking infrastructure per their security and compliance needs. You can expect new features to continue to optimize efficiency, performance, and security for software-defined datacenters.”
Server Core will also play a key role in the modernization of applications, Van Wesep emphasized. “One of our big pushes for next year is going to be around getting folks that have traditional .NET applications to drop those into containers running on Windows Server 2016, potentially even moving them into Azure,” he said. The new features will only be available to those opting for the new semi-annual channel, which will require Microsoft Software Assurance or Azure cloud subscriptions.
Microsoft explained how the new semi-annual channel release update will work. The company will offer new feature updates every spring and fall, the same model it recently moved to for Windows 10, Office 365 ProPlus and System Center. Microsoft will offer Windows Server previews shortly before the final release via its Windows Insiders for Business program, which is now open to those who want to sign up. Each semi-annual release will come with a pilot availability period of three to four months, and once the software is deemed stable, Microsoft will lock it down into a “Broad” release. Those releases will carry the Windows Server name with no year attached to it, instead of using the Windows 10 versioning model. The first release, scheduled for September, will be called Windows Server 1709. Chapple noted that the semi-annual channel feature updates are cumulative, containing previous updates. The semi-annual channel feature updates will get combined for Microsoft’s long-term servicing channel releases, a servicing model conceived for environments that can’t tolerate change.
The long-term servicing channel will include five years of mainstream support, five years of extended support and the option for six years of Premium Assurance. Van Wesep acknowledged that the long-term channel will be the most common in the near term. “I don’t imagine that the vast majority of the people will come out of the gates and say this is our new model and we will wholeheartedly switch to this — that would be naïve,” he said. “I think there has been enough demand and feedback from customers that they want a more active way of consuming our software that there will certainly be a meaningful size of the installed base that will start looking at this and working to adopt it. I can see a scenario where every customer would find both channels compelling for different parts of their organization.”
Indeed, many organizations may be resistant to such a model, and Van Wesep acknowledged that many existing applications and systems don’t lend themselves to a continuous release update model. But as many organizations look to transform their business processes or models over time, that can change. “This is on us to do the education process,” Van Wesep said. “People need to start thinking about the layers of abstraction between the app, the OS, and the underlying hypervisor/fabric. It all can be thought of independently and should be.”