A data lakehouse is essentially the next breed of cloud data lake and warehousing architecture that combines the best of both worlds. It is an architectural approach for managing all data formats (structured, semi-structured, or unstructured) as well as supporting multiple data workloads (data warehouse, BI, AI/ML, and streaming). Data lakehouses are underpinned by a new open system architecture that allows data teams to implement data structures through smart data management features similar to data warehouses over a low-cost storage platform that is similar to the ones used in data lakes. ... A data lakehouse architecture allows data teams to glean insights faster as they have the opportunity to harness data without accessing multiple systems. A data lakehouse architecture can also help companies ensure that data teams have the most accurate and updated data at their disposal for mission-critical machine learning, enterprise analytics initiatives, and reporting purposes. There are several reasons to look at modern data lakehouse architecture in order to drive sustainable data management practices.
When you get a chance to speak with executives, you typically don’t have much time to discuss details. And frankly, that’s not what executives are looking for, anyway. It’s important to phrase cybersecurity conversations in a way that resonates with the leaders. Messaging starts with understanding the C-suite and boards’ priorities. Usually, they are interested in big picture initiatives, so explain why cyber investment is critical to the success of these initiatives. For example, if the CEO wants to increase total revenue by 5% in the next year, explain how they can prevent major unnecessary losses from a cyber attack with an investment in cybersecurity. Once you know the executive team and board’s goals, look to specific members, and identify a potential ally. Has one team recently had a workplace security breach? Does one leader have a difficult time getting his or her team to understand the makings of a phishing scheme? These interests and experiences can help guide the explanation of the security solution. If you’re a CISO, you’re well-versed in cybersecurity, but remember that not everyone is as involved in the subject as you are, and business leaders probably will not understand technical jargon.
A bare metal machine is a dedicated server using dedicated hardware. Data centers have many bare metal servers that are racked and stacked in clusters, all interconnected through switches and routers. Human and automated users of a data center access the machines through access servers, high security firewalls and load balancers. The virtual machine introduced an operating system simulation layer between the bare metal server’s operating system and the application, so one bare metal server can support more than one application stack with a variety of operating systems. This provides a layer of abstraction that allows the servers in a data center to be software-configured and repurposed on demand. In this way, a virtual machine can be scaled horizontally, by configuring multiple parallel machines, or vertically, by configuring machines to allocate more power to a virtual machine. One of the problems with virtual machines is that the virtual operating system simulation layer is quite “thick,” and the time required to load and configure each VM typically takes some time. In a DevOps environment, changes occur frequently.
Many engineering teams rely on desktop products that only run on Microsoft Windows. Desktop engineering tools that perform tasks such as optical ray tracing, genome sequencing, or computational fluid dynamics often couple graphical user interfaces with complex algorithms that can take many hours to run on traditional workstations, even when powerful CPUs and large amounts of RAM are available. Until recently, there has been no convenient way to scale complex desktop computational engineering workloads seamlessly to the cloud. Fortunately, the advent of AWS Cloud Development Kit (CDK), AWS Elastic Container Service (ECS), and Docker finally make it easy to scale desktop engineering workloads written in C# and other languages to the cloud. ... The desktop component first builds and packages a Docker image that can perform the engineering workload (factor an integer). AWS CDK, executing on the desktop, deploys the Docker image to AWS and stands up cloud infrastructure consisting of input/output worker queues and a serverless ECS Fargate cluster.
Neuroscience also reveals why micromanaging is counterproductive. Donna Volpitta, an expert in “brain-based mental health literacy,” explained to me that the two most fundamental needs of the human brain are security and autonomy, both of which are built on trust. Leaders who instill a sense of trust in their employees foster that sense of security and autonomy and, in turn, loyalty. When leaders micromanage their employees, they undermine that sense of trust, which tends to breed evasion behaviors in employees. It’s a natural brain response. “Our brains have two basic operating modes—short-term and long-term,” Volpitta says. “Short-term is about survival. It’s the freeze-flight-fight response or, as I call it, the ‘grasshopper’ brain that is jumping all over. Long-term thinking considers consequences [and] relationships, and is necessary for complex problem solving. It’s the ‘ant’ brain, slower and steadier.” She says micromanagement constantly triggers short-term, survival thinking detrimental to both social interactions and task completion.
One of the key challenges of deep learning is the need for huge amounts of annotated data to train large neural networks. While this is the conventional way to train computer vision models, the latest generation of technology providers are taking an innovative approach that enables machine learning with comparatively less training data. This includes moving away from supervised learning to self-supervised and weakly supervised learnings where data availability is less of an issue. This approach, also known as few shot learning, detects objects as well as new concepts with considerably less input data. In many cases the algorithm can be trained with as little as 20 images. ... Privacy remains a major concern in the AI sector. In most cases, a business must share its data assets with the AI provider via third party servers or platforms when training computer vision models. Under such arrangements there is always the risk that the third party could be hacked or even exploit valuable metadata for its own projects. As a result, we’re seeing the rise of Privacy Enhancing Computation, which enables data to be shared between different ecosystems in order to create value, while maintaining data confidentiality.
Supporting remote and hybrid work requires organizations to invest in and implement new technologies that facilitate needs such as remote access, secure file-sharing, real-time collaboration and videoconferencing. Businesses must also hire professionals to configure, implement and maintain these tools with an eye towards security – a primary concern here, as businesses of all sizes now live or die by the availability and integrity of their data. The increasing complexity of IT environments – many of which are now pressured to support bring-your-own-device (BYOD) policies – has only intensified the need for competent cybersecurity talent. It’s not surprising that the ongoing shortage of trained professionals makes it difficult for organizations to expand their business and adopt new technologies. Almost half a million cybersecurity jobs remain open in the U.S. alone, forcing businesses to compete aggressively to fill these roles. Yet, economic pressures make it particularly difficult for small-to mid-sized businesses (SBMs) to play this game. Most cannot hope to match the high salaries that large enterprises offer.
The likely expansion of OCR’s mission into protecting the confidentiality of SUD data comes as actions to enforce the HITECH Breach Notification and HIPAA Security Rule appear to be at a standstill. According to the data compiled by OCR, in 2021 there were more than 660 breaches of the unauthorized disclosure of unsecured PHI reported by HIPAA-covered entities and their business associates that compromised the health information of over 40 million people. A significant number of the breaches reported to OCR appear to show violations of the HIPAA standards due to late reporting and failure to adequately secure information systems or train workforce members on safeguarding PHI. In 2021, OCR announced settlements in two enforcement actions involving compliance with the HIPAA Security Rule standards. OCR has been mum on its approach to enforcement of the HIPAA breach and security rules. One explanation could be the impact being felt by the 5th Circuit Court of Appeals decision overturning an enforcement action against the University of Texas MD Anderson Cancer Center.
It is truly the age of large models. Each of these models is bigger and more advanced than the previous one. Take, for example, GPT-3 – when it was introduced in 2020, it was the largest language model trained on 175 billion parameters. Fast forward one year, and we already have the GLaM model, which is a trillion weight model. Transformer models like GPT-3 and GLaM are transforming natural language processing. We are having active conversations around these models making job roles like writers and even programmers obsolete. While these can be dismissed as speculations, for now, one cannot deny that these large language models have truly transformed the field of NLP. Could this innovation be extended to other fields – like computer vision? Can we have a GPT-3 moment for computer vision? OpenAI recently released GLIDE, a text-to-image generator, where the researchers applied guided diffusion to the problem of text conditional image synthesis. For GLIDE, the researchers trained a 3.5 billion parameter diffusion model that uses a text encoder. Next, they compared CLIP (Contrastive Language-Image Pre-training) guidance and classifier free guidance.
To achieve a good level of agility the systems supporting the organization also have to quickly react and change to the surrounding environment. Legacy systems place a constraint on agility since they are often difficult to change or provide inefficient support to business activities. This is not unusual, at the time of the system design there were perhaps technology constraints that no longer exist, or the system was designed for a particular way of working that is no longer relevant. Legacy Modernization changes or replaces legacy systems, making the organization more efficient and cost-effective. Not only can this optimize existing business processes, but it can open new business opportunities. Security is an important driver for Legacy Modernization. Cyber-attacks on organizations are common and become more sophisticated over time. The security of a system degrades over time, and legacy systems may no longer have the support or the technologies required to deter modern attack methods, this makes them an easy target for hackers. This represents a significant business risk to an organization.
Quote for the day:
"Leadership is particularly necessary to ensure ready acceptance of the unfamiliar and that which is contrary to tradition." -- Cyril Falls