Those cybersecurity threats have sky-high substantially in recent because criminals have built lucrative businesses from stealing data and nation-states have come to see cybercrime as an opportunity to acquire information, influence, and advantage over their rivals. This has made a path for potential catastrophic attacks such as the WannaCrypt ransomware campaign which was being displayed in recent headlines. This evolving threat landscape has begun to change the way customers view the cloud. “It was only a few years ago when most of my customer conversations started with, ‘I can’t go to the cloud because of security. It’s not possible,’” said Julia White, Microsoft’s corporate vice president for Azure and security. “And now I have people, more often than not, saying, ‘I need to go to the cloud because of security.’” It’s not an exaggeration to say that cloud computing is completely changing our society. It’s ending major industries such as the retail sector, enabling the type of mathematical computation that is uplifting an artificial intelligence revolution and even having a profound impact on how we communicate with friends, family, and colleagues.
As is often the case in technology, everything old is new again. Suddenly, says Li, everything in deep learning is coming back to the innovations of compilers back in the day. "Compilers had become irrelevant" in recent years, he said, an area of computer science viewed as largely settled. "But because of deep learning, the compiler is coming back," he said. "We are in the middle of that transition." In his his PhD dissertation at Cornell, Li developed a computer framework for processing code in very large systems with what are called "non-uniform memory access," or NUMA. His program refashioned code loops for the most amount of parallel processing possible. But it also did something else particularly important: it decided which code should run depending on which memories the code needed to access at any given time. Today, says Li, deep learning is approaching the point where those same problems dominate. Deep learning's potential is mostly gated not by how many matrix multiplications can be computed but by how efficiently the program can access memory and bandwidth.
Event streaming employs the pub-sub approach to enable more accessible communication between systems. In the pub-sub architectural pattern, consumers subscribe to a topic or event, and producers post to these topics for consumers’ consumption. The pub-sub design decouples the publisher and subscriber systems, making it easier to scale each system individually. The publisher and subscriber systems communicate through a message broker like Apache Pulsar. When a state changes or an event occurs, the producer sends the data (data sources include web apps, social media and IoT devices) to the broker, after which the broker relates the event to the subscriber, who then consumes the event. Event streaming involves the continuous flow of data from sources like applications, databases, sensors and IoT devices. Event streams employ stream processing, in which data undergoes processing and analysis during generation. This quick processing translates to faster results, which is valuable for businesses with a limited time window for taking action, as with any real-time application.
In a nutshell, the changes that come into effect from October allow customers with Software Assurance or subscription licenses to use these existing licenses "to install software on any outsourcers' infrastructure" of their choice. But as The Register noted at the time, this specifically excludes "Listed Providers", a group that just happens to include Microsoft's biggest cloud rivals – AWS, Google and Alibaba – as well as Microsoft's own Azure cloud, in a bid to steer customers to Microsoft's partner network. ... These criticisms are not entirely new, and some in the cloud sector made similar points following Microsoft's disclosure of some of the licensing changes it intended to make back in May. One cloud operator who requested anonymity told The Register in June that Redmond's proposed changes fail to "move the needle" and ignore the company's "other problematic practices." Another AWS exec, Matt Garman, posted on LinkedIn in July that Microsoft's proposed changes did not represent fair licensing practice and were not what customers wanted.
Built on 16nm technology, the MLSoC’s processing system consists of computer vision processors for image pre- and post-processing, coupled with dedicated ML acceleration and high-performance application processors. Surrounding the real-time intelligent video processing are memory interfaces, communication interfaces, and system management — all connected via a network-on-chip (NoC). The MLSoC features low operating power and high ML processing capacity, making it ideal as a standalone edge-based system controller, or to add an ML-offload accelerator for processors, ASICs and other devices. The software-first approach includes carefully-defined intermediate representations (including the TVM Relay IR), along with novel compiler-optimization techniques. ... Many ML startups are focused on building only pure ML accelerators and not an SoC that has a computer-vision processor, applications processors, CODECs, and external memory interfaces that enable the MLSoC to be used as a stand-alone solution not needing to connect to a host processor. Other solutions usually lack network flexibility, performance per watt, and push-button efficiency – all of which are required to make ML effortless for the embedded edge.
“Now more than ever, we’re seeing a pressing demand for CIOs to deliver digital transformation that enables business growth to energize the top line or optimize operations to eliminate cost and help the bottom line,” says Savio Lobo, CIO of Ensono. This requires the CIO to have a deep understanding of the business and surface decisions that may influence these objectives. Large-scale digital solutions and capabilities, however, often cannot be implemented simultaneously, especially when they require significant change in how customers and staff engage with people and processes. This means ruthless prioritization decisions may need to be made that include what is moving forward at any given time and equally importantly, what is not. “While executing a large initiative, there will also be people, process and technology choices to be made and these need to be made in a timely manner,” Lobo adds. This may look unique for every organization but should include collaboration on the discovery and implementation and an open feedback loop for how systems and processes are working or not working in each stakeholder’s favor.
A ‘Trusted Computing’ model, like the one developed by the Trusted Computing Group (TCG), can be easily applied to all four of these AI elements in order to fully secure a rogue AI. Considering the data set element of an AI, a Trusted Platform Module (TPM) can be used to sign and verify that data has come from a trusted source. A hardware Root of Trust, such as the Device Identifier Composition Engine (DICE), can make sure that sensors and other connected devices maintain high levels of integrity and continue to provide accurate data. Boot layers within a system each receive a DICE secret, which combines the preceding secret on the previous layer with the measurement of the current one. This ensures that when there is a successful exploit, the exposed layer’s measurements and secrets will be different, securing data and protecting itself from any data disclosure. DICE also automatically re-keys the device if a flaw is unearthed within the device firmware. The strong attestation offered by the hardware makes it a great tool to discover any vulnerabilities in any required updates.
The practical implication for software engineering management is to first address feedback loops that generate a lot of bugs/issues to get your capacity back. For example, if you have a fragile architecture or code of low maintainability that requires a lot of rework after any new change implementation, it is obvious that refactoring is necessary to regain engineering productivity; otherwise, engineering team capacity will be low. The last observation is that the lead time will depend on the simulation duration, the longer you run the value stream, the higher the number of lead times variants you will get. Such behavior is the direct implication of the value stream structure with the redo feedback loop and its probability distribution between the output queue and the redo queue. If you are an engineering manager who inherited legacy code with significant accumulated debt, it might be reasonable to consider incremental solution rewriting. Otherwise, the speed of delivery will be very slow forever, not only for the modernization time. The art of simplicity; greater complexity yields more variations which increase the probability of results occurring outside of acceptable parameters.
Realizing the benefits of edge computing depends on a thoughtful strategy and careful evaluation of your use cases, in part to ensure that the upside will dwarf the natural complexity of edge environments. “CIOs shouldn’t adopt or force edge computing just because it’s the trendy thing – there are real problems that it’s intended to solve, and not all scenarios have those problems,” says Jeremy Linden, senior director of product management at Asimily. Part of the intrinsic challenge here is that one of edge computing’s biggest problem-solution fits – latency – has sweeping appeal. Not many IT leaders are pining for slower applications. But that doesn’t mean it’s a good idea (or even feasible) to move everything out of your datacenter or cloud to the edge. “So for example, an autonomous car may have some of the workload in the cloud, but it inherently needs to react to events very quickly (to avoid danger) and do so in situations where internet connectivity may not be available,” Linden says. “This is a scenario where edge computing makes sense.” In Linden’s own work – Asimily does IoT security for healthcare and medical devices – optimizing the cost-benefit evaluation requires a granular look at workloads.
Tenable wants to provide customers with more context around what threat actors are exploiting in the wild to both refine and leverage the analytics capabilities the company has honed, Yoran says. Tenable must have context around what's mission-critical in a customer's organization to help clients truly understand their risk and exposure rather than just add to their cyber noise, he adds. Tenable has spent more on vulnerability management-focused R&D over the past half-decade than its two closest competitors combined, which has allowed the firm to deliver differentiated capabilities, Yoran says. Unlike competitors who have expanded their offerings to include everything from logging and SIEM to EDR and managed security services, Yoran says Tenable has remained laser-focused on risk. "The three primary vulnerability management vendors have three very different strategies and they've been on divergent paths for a long time," Yoran says. "For us, the key to success has been and will continue to be that focus on helping people assess and understand risk."
Quote for the day:
"Get your facts first, then you can distort them as you please." -- Mark Twain