Daily Tech Digest - November 03, 2020

Why Securing Secrets in Cloud and Container Environments Is Important – and How to Do It

In containerized environments, secrets auditing tools make it possible to recognize the presence of secrets within source code repositories, container images, across CI/CD pipelines, and beyond. Deploying container services will activate platform and orchestrator security measures that distribute, encrypt and properly manage secrets. By default, secrets are secured in system containers or services — and this protection suffices in most use cases. However, for especially sensitive workloads — and Uber’s customer database backend service is a strong example, as are any data encryption or standard image scanning use cases — it’s not adequate to simply rely on conventional secret store security and secret distribution. These sensitive use cases call for more robust defense in depth protections. Within container environments, defense-in-depth implementations leverage deep packet inspection (DPI) and data leakage prevention (DLP) to enable secrets monitoring while they’re being used. Any transmission of a secret via network packets can be recognized, flagged and blocked if inappropriate. In this way, the most sensitive data can be effectively secured throughout the full container lifecycle, and attacks that could otherwise result in breach incidents can be thwarted due to this additional layer of safeguards.

Large-Scale Multilingual AI Models from Google, Facebook, and Microsoft

While Google's and Microsoft's models are designed to be fine-tuned for NLP tasks such as question-answering, Facebook has focused on the problem of neural machine translation (NMT). Again, these models are often trained on publicly-available data, consisting of "parallel" texts in two different languages, and again the problem of low-resource languages is common. Most models therefore train on data where one of the languages is English, and although the resulting models can do a "zero-shot" translation between two non-English languages, often the quality of such translations is sub-par. To address this problem, Facebook's researchers first collected a dataset of parallel texts by mining Common Crawl data for "sentences that could be potential translations," mapping sentences into an embedding space using an existing deep-learning model called LASER and finding pairs of sentences from different languages with similar embedding values. The team trained a Transformer model of 15.4B parameters on this data. The resulting model can translate between 100 languages without "pivoting" through English, with performance comparable to dedicated bi-lingual models.

How to Prevent Pwned and Reused Passwords in Your Active Directory

There are many different types of dangerous passwords that can expose your organization to tremendous risk. One way that cybercriminals compromise environments is by making use of breached password data. This allows launching password spraying attacks on your environment. Password spraying involves trying only a few passwords against a large number of end-users. In a password spraying attack, cybercriminals will often use databases of breached passwords, a.k.a pwned passwords, to effectively try these passwords against user accounts in your environment. The philosophy here is that across many different organizations, users tend to think in very similar ways when it comes to creating passwords they can remember. Often passwords exposed in other breaches will be passwords that other users are using in totally different environments. This, of course, increases risk since any compromise of the password will expose not a single account but multiple accounts if used across different systems. Pwned passwords are dangerous and can expose your organization to the risks of compromise, ransomware, and data breach threats. What types of tools are available to help discover and mitigate these types of password risks in your environment?

The practice of DevOps

Continuous integration (CI) is the process that aligns the code and build phases in the DevOps pipeline. This is the process where new code is merged with the existing structure, and engineers ensure that everything is working fine. The developers who make frequent changes to the code, update the same on the shared central code repository. The repository starts with a master branch, which is a long-term stable branch. For every new feature, a new branch is created, and the developer regularly (daily) commits his code to this branch. After the development for a feature is complete, a pull request is created to the release branch. Similarly, a pull request is created to the master branch and the code is merged. We have seen slight variations to these practices across organisations. Sometimes the developers maintain a fork, or copy, of the central repository. This limits the merge issues to their own fork and isolates the central repository from the risk of corruption. Sometimes, the new branches don’t branch out from the feature branch but from the release or master branch. Small and mid-size companies often use open sites like GitHub for their code repository, while larger firms use Bitbucket as their code repository, which is not free.

The 5 Biggest Cloud Computing Trends In 2021

Currently, the big public cloud providers - Amazon, Microsoft, Google, and so on – take something of a walled garden approach to the services they provide. And why not? Their business model has involved promoting their platforms as one-stop-shops, covering all of an organization's cloud, data, and compute requirements. In practice, however, industry is increasingly turning to hybrid or multi-cloud environments (see below), with requirements for infrastructure to be deployed across multiple models. ... As far as cloud goes, AI is a key enabler of several ways in which we can expect technology to adapt to our needs throughout 2021. Cloud-based as-a-service platforms enable users on just about any budget and with any level of skill to access machine learning functions such as image recognition tools, language processing, and recommendation engines. Cloud will continue to allow these revolutionary toolsets to become more widely deployed by enterprises of all sizes and in all fields, leading to increased productivity and efficiency. ... Amazon most recently joined the ranks of tech giants and startups offering their own platform for cloud gaming. Just as with music and video streaming before it, cloud gaming promises to revolutionize the way we consume entertainment media by offering instant access to vast libraries of games that can be played for a monthly subscription.

Quantum computers are coming. Get ready for them to change everything

The challenge lies in building quantum computers that contain enough qubits for useful calculations to be carried out. Qubits are temperamental: they are error-prone, hard to control, and always on the verge of falling out of their quantum state. Typically, scientists have to encase quantum computers in extremely cold, large-scale refrigerators, just to make sure that qubits remain stable. That's impractical, to say the least. This is, in essence, why quantum computing is still in its infancy. Most quantum computers currently work with less than 100 qubits, and tech giants such as IBM and Google are racing to increase that number in order to build a meaningful quantum computer as early as possible. Recently, IBM ambitiously unveiled a roadmap to a million-qubit system, and said that it expects a fault-tolerant quantum computer to be an achievable goal during the next ten years. Although it's early days for quantum computing, there is still plenty of interest from businesses willing to experiment with what could prove to be a significant development. "Multiple companies are conducting learning experiments to help quantum computing move from the experimentation phase to commercial use at scale," Ivan Ostojic, partner at consultant McKinsey, tells ZDNet.

How enterprise architects and software architects can better collaborate

After the EAs have developed the enterprise architecture map, they should share these plans with software architects across each solution, application, or system. After all, it’s important that the software architect who works most closely with the solution shares their own insight and clarity with the enterprise architecture, who is more concerned with high-level architecture. Software architects and EAs can collaborate and suggest changes or improvements based on the existing architecture. Software architects can then go in and map out the new architecture based on the business requirements. Non-technical leaders can gain a better understanding and this can lead to quicker alignment. Software architects can evaluate the quality of architecture and share their learnings with the enterprise architect, who can then incorporate findings into the enterprise architecture. Software architects are also largely interested in standardization and they can help enterprise architects scale that standardization across the business. Once the EA has developed a full model or map, it’s easier to see where assets can be reused. Software architects can recommend standardization and innovation and weigh in on the EA’s suggestions of optimizing enterprise resources.

Dealing with Psychopaths and Narcissists during Agile Change

Many of the techniques or practices you use with healthy people do not work well with psychopaths or narcissists. For example, if you are using the Scrum framework, it is very risky to include a toxic person as part of a retrospective meeting. Countless consultants also believe that coaching works with most folks. However, the psychopathic person normally ends up learning the coach’s tools and manipulating him or her for their own purpose. This obviously aggravates the problem. ... From an organizational point of view, these toxic people are excellent professionals, because they look like they perform almost any task successfully. This helps a company to “tick off” necessary accomplishments in the short-term to increase agile maturity, managers to get their bonuses, and the psychopath to obtain greater prestige. Obviously, these things are not sustainable, and what seems to be agility is transformed in the medium term into fragility and loss of resilience. Agile also requires—apart from good organizational health—execution with purpose and visions and goals that involve feelings and inspire people to move forward.

Responsible technology: how can the tech industry become more ethical?

Three priorities right now should be crafting smart regulations, increasing the diversity of thought in the tech industry (i.e. adding more social scientists), and bridging the gulf between those that develop the technology and those that are most impacted by its deployment. If we truly want to impact behavior, smart regulations can be an effective tool. Right now, there is often a tension between “being ethical” and being successful from a business perspective. In particular, social media platforms typically rely on an ad-based business model where the interests of advertisers can run counter to the interests of users. Adding diverse thinkers to the tech industry is important because “tech” should not be confined to technologists. What we are developing and deploying is impacting people, which heightens the need to incorporate disciplines that naturally understand human behavior. By bridging the gulf between the individuals developing and deploying technology and those impacted by it, we can better align our technology with our individual and societal needs. Facial recognition is a prime example of a technology that is being deployed faster than our ability to understand its impact on communities.

Enterprise architecture has only begun to tap the power of digital platforms

The problem that enterprises have been encountering, Ross, says, is getting hung up at stage three. “We observed massive failures in business transformations that frankly were lasting six, eight, 10 years. It’s so hard because it’s an exercise in reductionism, in tight focus,” she said. “We recommend that companies zero in on their single most important data. this is the packaged data you keep – the customer data, the supply chain data… this is the thing that matters most. If they get this right, then things will take off.” The challenge is now moving past this stage, as in the fourth stage, “we actually understand now that what’s starting to happen is we can start to componentize our business,” says Ross. She estimates that only about seven percent of companies have reached this stage. “This is not just about plugging modules into this platform, this is about recognizing that any product or process can be decomposed into people, process and technology bundles. And we can assign individual teams or even individuals’ accountability for one manageable piece that that team can keep up to date, improve with new technology, and respond to customer demand.”

Quote for the day:

"Let him who would be moved to convince others, be first moved to convince himself." -- Thomas Carlyle

1 comment:

  1. Copywriting AI tools often employ techniques such as sentiment analysis, tone detection, and content optimization to ensure that the generated copy aligns with the desired brand voice, resonates with the target audience, and maximizes conversion rates.