Trend 2: Cybersecurity Mesh
Everyone knows that the number and nature of cyberthreats is increasing. We also know that there are no perfect solutions – and never will be. We know that most companies do not fully understand the threats and therefore underspend in cybersecurity or spend the wrong way. Statements like these are what you hear from vendors:
“Cybersecurity mesh is a flexible, composable architecture that integrates widely distributed and disparate security services.
“Cybersecurity mesh enables best-of-breed, stand-alone security solutions to work together to improve overall security while moving control points closer to the assets they’re designed to protect.
“It can quickly and reliably verify identity, context and policy adherence across cloud and noncloud environments”
These are normative and prescriptive statements, not actionable ones. Of course we want “composable architecture that integrates widely distributed and disparate security services.” Who doesn’t? There’s always an ocean between what we want and how to get it. This a not trend, just an old aspiration.
AI on the rise
Related to autonomic and hyperautomation growth will be the increased use of AI and in particular, generative AI—machine learning methods that learn about content or objects from their data, and use it to generate brand-new, completely original content, Groombridge said.
Generative AI can be used for a range of activities such as creating software code, facilitating drug development, and targeting marketing. It could also be misused for scams, fraud, political disinformation, and forged identities. By 2025, Gartner expects generative AI to account for 10% of all data produced, up from less than 1% today, Groombridge said.
Можно ли все этоспрогнозировать?
Предполагать, кто будет нужен рынку через пять, 10 и более лет, можно: этим как раз и занимаются футурологи. Свои прогнозы они строят не на субъективном восприятии, а на больших массивах данных. Аналитикам сегодня помогают технологии на основе искусственного интеллекта, которые обрабатывают всю доступную информацию: об экономической ситуации, научных и технических достижениях, спросе на рынке труда.
На основе этих данных составляются списки или целые «атласы» профессий будущего: например, «Атлас новых профессий» «Сколково» или НИУ ВШЭ и «Сбера». Также можно заглянуть в Национальную программу «Цифровая экономика Российской Федерации» — она косвенно дает понять, какие специалисты понадобятся для реализации этой программы.
Но с абсолютной точностью сказать, что через 10 лет
будет страшно востребована вот эта десятка специальностей, невозможно. Прогнозы полезны для понимания общей тенденции, но воспринимать их буквально не стоит.
Дина Муштанова
Президент Международной Ассоциации экспертов по профориентации
«Ради интереса посмотрите, какие предсказания относительно профессий будущего были 5–10 лет назад. Консультант по роботам. Сити-фермер. Архитектор живых систем. Эти и многие другие профессии, распространение которых предсказывали эксперты, востребованности на рынке пока не получили. В то же время другие направления — например, зерокодинг или разработка моделей Big Data — действительно пользуются большим спросом»
Ставка на технологичность процессов совершенно точно будет трендом в большинстве сфер. Это значит, что рынку потребуется все больше специалистов в IT-секторе.
Что нужно учитывать при реализации проектов на основе глубокого обучения
Реализация проектов на основе глубокого обучения имеет ряд особенностей. Прежде всего нужно понимать, что это не проекты одного дня. После стандартных для проектов машинного зрения этапов выбора оптической схемы, сбора и разметки данных следует этап выбора архитектуры нейронной сети и последующий итерационный процесс ее обучения. Часто невозможно заранее предсказать, какие показатели качества (ошибки пропуска/ложной тревоги или критические/некритические ошибки классификации) будут достигнуты в реальных условиях производства. Как следствие, в проектах машинного зрения на основе глубокого обучения на начальном этапе присутствуют научно-технические риски, которые должны разделить между собой заказчик и разработчик.
Далеко не все предприятия готовы инвестировать в инновационные разработки и пилотные проекты, предпочитая получать готовое решение «под ключ». Такой подход заказчика в отношении, по сути, уникальных задач, требующих применения искусственного интеллекта для разработки эксклюзивного продукта, как показывает опыт, малоэффективен. Нежелание руководства предприятия инвестировать в разработку может обернуться тем, что поиск исполнителя для реализации проекта затянется на годы и решить задачу вообще не удастся.
Несколько лет назад мы разработали для компании «АЛРОСА» две сложные системы автоматической сортировки алмазов по цвету и форме. Объект контроля — неграненые алмазы диаметром 1–5 мм, подаваемые в зону контроля со скоростью 20 шт./с. Камеры машинного зрения выполняют съемку алмаза при его движении и в свободном падении. Полученные изображения передаются на сервер, программное обеспечение анализирует характеристики алмаза на изображениях и принимает решение о его отнесении к определенному классу (рис. 2). Все это происходит практически мгновенно, пока алмаз пролетает дистанцию всего в 30 см, после чего на основании выданного системой результата он отсекается в соответствующий накопительный бункер. На решение этой задачи было потрачено в общей сложности четыре года — от начального этапа проведения эксперимента в формате НИОКР, создания опытно-промышленного образца и до разработки промышленной системы. Срок достаточно долгий, но заказчик прекрасно понимал, что разработка такой уникальной системы должна вестись поэтапно, чтобы минимизировать возможные риски для обеих сторон.
Рис. 2. Выборка данных для обучения модели классификации алмазов по цвету
Final Thoughts
The main problem with Gartner’s (and other) trends lists is they cannot help but be repetitive and overlapping, especially with prior years. While things move fast in technology land, major trends – like applications architecture and AI – will be “trends” for years. “Updates” are interesting but not that useful. Sure, more people are thinking about microservices and containers than they were last year, and more companies are investing in AI pilots. It’s hard to describe truly new trends, so technology prognosticators rename old ones with catchy names. But some of the names are just too obvious. Maybe next year the trends will be new or the (re) names more creative. Or maybe they’ll just be called aspirations.
Ситуация с производителями оборудования:
- с их шлемами Oculus Quest сделал ставку как раз на гибридные устройства, пишут аналитики. По прогнозу в 2020 году компания продаст порядка 1,2 гарнитур из линейки Quest, а в 2025 году — 5,6 млн гарнитур.
- Однако за эти пять лет Facebook не сможет окончательно укрепить свои позиции в нише. Omdia ожидает, что в ближайшее время на рынке появятся более доступные гибридные шлемы (преимущественно из Китая). На фоне этого доля Facebook в глобальных продажах автономного VR снизится с 48% в 2020 году до 35% в 2025 году.
- Недавно Sony объявила, что не станет запускать PSVR 2 в следующем году. Omdia предполагает, что эта консольная VR выйдет в конце 2022 года, когда компании уже не надо будет так сильно сосредотачиваться на своей последней новинке PS5.
- Вероятно, в ожидании старта PSVR 2 игроки будут реже приобретать предыдущую модель гарнитуры. Зато в премьерные три месяца тираж PSVR 2 достигнет отметки в 900 тыс. проданных устройств.
- Аналитики не делают прогнозов для VR-устройств Valve, но подчеркивают — гарнитура Valve Index, подключаемая к PC, пока остается популярной. Этому во многом способствовал релиз Half Life: Alyx. Однако высокая цена ($999) и необходимость в мощном компьютере отпугнули значительную часть игроков.
- Тем временем успешное будущее шлемов HTC под угрозой, отмечает Omdia. Эксперименты с расширением портфеля и запуском проектов вроде Vive Cosmos могут выйти компании боком.
- Из-за медленных темпов роста VR в скором времени может пострадать и практически весь интерактивный видеоконтент с обзором в 360 градусов. Речь идет о BBC VR Hub, Google Spotlight Stories, Jaunt и Oculus Story Studio.
Также по теме:
- VR-хоррор Phasmophobia вошел в топ-10 бестселлеров в Steam
- Японский VR-разработчик Thirdverse привлек $8,5 млн инвестиций
- Авторы VR-боевика The Walking Dead: Onslaught привлекли $16,7 млн
Institutionalize Trust
One of the key pillars of data-driven culture is the trust on data and analytics as an effective business lever among the associates in the organization. The following can be focus areas in relation to institutionalizing trust on data analytics capabilities across the organization:
- Connected governance: The goal is to have governance within the business unit to address the operational challenges in a timely manner and also, connected business units and changing external market. Data analytics can be leveraged to gather and analyze the data from different business units and market intelligence.
- AI risk management
- Vendor and regional ecosystems
- Expansion to the edge
Trend 1: Data Fabric
Data has always been essential to operational and strategic effectiveness. “Making data available everywhere” has been a priority for decades. Accessing the right data at the right time across multiple platforms and applications has always been a problem. “Analytics” is how we decided to brand both the problems and solutions. Old problems; new names. While “data fabric” is cute, it’s just the latest way we describe data nirvana, but in a proprietary technology world, it remains a pipedream.
MORE FROMFORBES ADVISOR
Income Tax Calculator: Estimate Your Taxes
Should we continue to clean, integrate and present structured and unstructured data? Of course, through whatever cost-effective means we can. But we should also recognize there are limits to what we can cost-effectively achieve, especially when we pursue best of breed data and applications strategies? Gartner also argues that data fabrics “can reduce data management efforts by up to 70%.” How in the world is this a knowable number?
Общие данные из прогноза:
- За 2020 год пользователи купят 6,4 млн VR-гарнитур, причем 3,3 млн из этих устройств будут автономными (не требующими для работы PC или консоли).
- Выручка от продажи VR-контента к концу 2020 года достигнет $1,1 млрд, а общий оборот сегмента составит $3,2 млрд.
- Вот совокупная выручка рынка виртуальной реальности в 2025 году составит около $10 млрд. Чуть меньше половины суммы ($4 млрд) принесет VR-контент, в основном это будут игры (90%).
- К 2025 году объем проданных гарнитур вырастет уже до 45 млн, но сам сегмент VR так и останется нишевым.
- Спустя пять лет VR-гарнитуры появятся в домах только 3% жителей из 32 ведущих стран (какие именно страны имеются в виду — не уточняется). Для сравнения, сейчас шлемами виртуальной реальности пользуются не более 1,2% населения.
- VR будет расти и дальше, но не очень быстро. По мнению аналитиков, до массового внедрения этой технологии потребуется не менее десяти лет.
- При этом пользователи будут все сильнее ценить гибридную автономность гарнитур, считает Omdia. Именно на эту категорию придется большая часть продаж в 2025 году.
Прогноз покрытия сетей 5G и «последняя миля»
| Частоты | Ширина полосы | Сценарии | Характеристика |
|---|---|---|---|
| выше 7 ГГц (FR2) | 800 МГц | eMBB | Сверхвысокая скорость, маленькое покрытие и только на улицах |
| 2 ГГц…7 ГГц (FR1) | 100 МГц | eMBB, URLLC, mMTC | Высокая скорость, широкое покрытие на улицах, удовлетворительное покрытие в помещениях |
| < 2 ГГц (FR1) | 20 МГц | eMBB, URLLC, mMTC | Средняя скорость, вездесущее покрытие на улицах и в помещениях |
Обобщенная схема покрытия сетями 2G, LTE и 5G до 2025 годаFWA и проводные/оптоволоконные подключения на «последней миле»CPE внутри дома (Indoor CPE) и настенный (Outdoor CPE)
Trends Redux
Let’s pretend that Gartner invited us to describe the trends we see. What would they look like? Not what should they look like, but what they actually are. First, let’s reorganize the Gartner trends into baskets and then assess how well they’re doing out there.
Three baskets jump out:
Automation
- Decision Intelligence
- Hyperautomation
- AI Engineering
- Autonomic Systems
- Generative AI
Infrastructure
- Data Fabric
- Cybersecurity Mesh
- Privacy-Enhancing Computation
- Cloud-Native Platforms
Applications
- Composable Applications
- Distributed Enterprises
- Total Experience
So what’s happening out there that might be described as trends? Survey data suggests that there’s still some skepticism about how quickly executives believe they have to ramp up their AI and machine learning (ML) investments. That said, there’s a steady increase in AI/ML pilots. So that’s a trend that will accelerate over time so long as impact is quantifiable. We do know that trends in RPA are positive. We know that there are new “Centers of Excellence” being created and we also know that the adoption of AI/ML is vertically driven: some industries are piloting applications more than others. There’s also a rank-ordering of AI/ML applications, like chatbots (“conversational AI”) and selection applications (such as selecting the best/worse candidates for loans, “admission,” and any binary decision). We know that adoption rates are dependent upon strategy, which has yet to fully embrace intelligent systems as the future that will define profitable growth. Deloitte reports that the distribution across “path seekers,” “transformers,” “starters” and “underachievers” in AI is challenging, suggesting that more than two thirds of companies are not “high achievers.” Deloitte also reports that companies are bullish on AI/ML, that eventually they expect AI/ML to really impact their business. So the real trend is cautious optimism.
Data trends? Analytics is now mainstream – at least by name and intent. But clean, integrated, accessible data remains elusive. Why? The old standbys, such as competing data formats across applications, “lost” data, bad data, duplicate data and no enterprise data architecture, are among the problems that have haunted us for decades. So seamless, integrated and accessible data is an ongoing aspiration. But trend? Yes, if by trend we mean investment trends. Make no mistake this is an ongoing slog.
Applications architecture is finally getting some traction. Companies are moving toward microservices-based applications as they remove their legacy applications from life support. “Applications rationalization” is a real thing. It saves money and moves the ball toward the microservices goal line. Container technology has evolved nicely and, perhaps most importantly, the major cloud vendors offer a variety of solutions here. Perhaps this is the real trend. Edge computing is part of this, and total experience applications will always be what everyone strives to build.
Ключевые показатели стандарта 5G и технологии
- пиковая скорость передачи данных на линии вниз (Downlink) 20 Гбит/с (спектральная эффективность 30 бит/с/Гц);
- пиковая скорость передачи данных на линии вверх (Uplink) 10 Гбит/с (спектральная эффективность 15 бит/с/Гц);
- минимальная задержка в подсистеме радиодоступа для сервисов URLLC — 0,5 мс, для сервисов eMBB — 4 мс;
- максимальная плотность подключенных к сети в городских условиях устройств из мира IoT – 1’000’000 устройств/кв.км;
- автономная работа устройств из мира IoT без подзарядки аккумулятора в течение 10 лет;
- поддержка мобильности при максимальной скорости передвижения объектов 500 км/ч.
некоторых
Частота и ширина полос
| Блок радиочастот | Радиочастотный диапазон |
|---|---|
| FR1 | 450 MHz – 6 000 MHz |
| FR2 | 24’250 MHz – 52’600 MHz |
Massive MIMO и Beam Forming (формирование луча)
2D MIMO антенна (слева) и Massive MIMO антенна (справа)
- мощный сигнал на выходе в направлении к UE;
- сильный уровень сигнал/шум в направлении от UE;
- отсутствие межсотовой интерференции;
- значительное увеличение количества каналов связи на одну соту.
| Sub6G | mmWave | |
|---|---|---|
| Порядок MIMO | до 8х8 | 2х2 |
| Смысл | Статичное пространственное мультиплексирование для множества пользователей | Динамическое формирование луча для одного пользователя |
| Характеристика | Многолучевое распространение, идеален для пространственного мультиплексирования. Протяженная зона покрытия, покрытие внутри зданий. | Распространение в прямой видимости. Массовые соединения со сверх широкой полосой пропускания. |
Сценарии и примеры оказания услуг мобильной связи в сетях 5G
- eMBB (enhanced Mobile Broadband), сверхширокополосная мобильная связь;
- URLLC (Ultra-Reliable Low Latency Communication), сверхнадежная связь с низкими задержками;
- mMTC (Massive Machine-Type Communications), массовая межмашинная связь.
Три сценария оказания услуг мобильной связи
Augment People & Decisions
One of the key objective of data and analytics in current times is to deliver business value/impact by leveraging data, analytics and people expertise to make right decisions. The following are key focus areas in this relation:
- Context enriched analysis
- Business composed data analytics
- Decision centric data analytics
- Skills & literacy shortfall
Lets understand each of the above focus areas in detail.
- Context enriched analysis: Context enriched analysis is a data analytics approach that incorporates contextual information into the analysis in order to improve its accuracy and relevance. The idea is that by understanding the context in which data was collected, analysts can better interpret its meaning and identify patterns that would otherwise be hidden. Contextual information can include anything from the location of a data point to the time of day it was collected. Context enriched analysis is often used in fields such as marketing, where understanding customer behavior is essential to making effective decisions. According to Gartner, context enriched analysis will become increasingly important in the coming years as the volume and variety of data continue to grow.
- Business-composed data analytics: Business composed data analytics is the process of building packaged business capabilities (PBC) related to data and analytics with a goal to deliver valuable insights in an agile manner to improve business operations. Analytics packaged business capabilities is about delivering insights by leveraging packaged capabilities in form of traditional BI dashboards, self-service analytics, predictive modeling, prescriptive modeling, reports, etc. Data PBC is about building capabilities around getting quick access to data through data fabric layer. Data fabric layer is an abstract layer of software that hides the complexity of data management from users and applications. Data PBC will help you get to insights faster by leveraging existing analytics capabilities
- Decision-centric data analytics: Decision centric data analytics is an approach to data analytics that focuses on making better decisions, rather than simply understanding what has happened in the past. The idea is that by understanding the factors that influence decision-making, analysts can provide more accurate and relevant insights. Gartner estimates that by 2023, more than 33% of large organizations will have analysts practicing decision intelligence, including decision modeling. Decision intelligence can be defined as a set of processes and methods used to understand how humans make decisions, and how those decisions can be improved. The two key aspects of decision intelligence is how to ensure/monitor that good decisions are made, and, how the decisions are made. Data analytics play key role in both these aspects. Here is a great video on how decision intelligence can impact business outcomes.
- Skills & literacy shortage: Skills and data literacy shortfall is a Gartner identified Top 12 Data and Analytics Strategy Trend for 2022. As per Gartner, by 2024, 50% of enterprises will have a data literacy program, compared with 20% in 2020. The skills shortage is not just about finding people with the right technical skills, but also about finding people who understand how to use data to make decisions. Data literacy is the ability to read, write and communicate data in a way that it can be understood by others. It also includes the ability to understand and use statistical techniques to analyze data. Gartner predicts that the demand for data literacy will continue to grow in the coming years as organizations increasingly rely on data to make decisions.
AI engineers needed
As the use of AI grows so does the need for AI engineering, a discipline focused on the governance and life-cycle management of a wide range of operationalized AI and decision models, such as machine learning and knowledge graphs.
“For fusion teams working on AI, the real differentiator for their organizations will lie in their ability to continually enhance value through rapid AI change,” Groombridge said. “By 2025, the 10% of enterprises that establish AI-engineering best practices will generate at least three times more value from their AI efforts than the 90% of enterprises that do not.”
Other trends Gartner says will affect the IT environment in 2022 include:
- Cloud-Native Platforms (CNPs): The notion of lifting and shifting legacy applications to cloud environments doesn’t really play well in the real world, Groombridge said. Rather, a cloud native architecture rebuilds applications to produce highly automated cloud services that deliver digital capabilities everywhere and anywhere. CNPs use the core capabilities of cloud computing to provide scalable and elastic IT-related capabilities as a service to technology creators. For this reason, Gartner predicts that cloud-native platforms will serve as the foundation for more than 95% of new digital initiatives by 2025, up from less than 40% in 2021.
- Composable Applications: Composable application architectures allow for the quick change or breakout from an existing app to another to address a particular business need. Composable application architecture enables this adaptability, and those that have adopted it will outpace competition by 80% in the speed of implementing new features, Groombridge said.
- Privacy-Enhancing Computation (PEC): PEC techniques protect personal and sensitive information at the data, software, and hardware level, making it possible to securely share, pool, and analyze data without compromising confidentiality or privacy. Gartner said it expects 60% of large organizations to use one or more PEC techniques by 2025.
- Decision Intelligence: Decision intelligence is a discipline used to improve decision making by explicitly understanding and engineering how decisions are made, and how outcomes are evaluated, managed, and improved by feedback. Gartner predicts that in the next two years, a third of large organizations will use decision intelligence to improve competitive advantage.
- Total Experience (TX): TX is a business strategy that combines the disciplines of customer experience (CX), employee experience (EX), user experience (UX) and multiexperience (MX). The goal of TX is to drive greater customer and employee confidence, satisfaction, loyalty and advocacy. Organizations will increase revenue and profit by achieving adaptive and resilient TX business outcomes.
Gartner Research Report Identifies Eight Categories of Data Integration Technology
Gartner reports that 75% of its clients now use more than one cloud service provider, creating a need for data integration tools that can bring together data from increasingly complex and distributed cloud environments. Gartner’s new Implementing Multicloud Data Integration report was designed to provide D&A leaders with guidance on cloud data integration by rigorously categorizing and analyzing the available data integration technologies in the marketplace.
As part of its analysis, Gartner defines four patterns of data integration (Data Ingestion, Data Consistency, Multistep Process Integration, and Composite Services) and three data integration styles (data-centric, event-centric, and application-centric), then maps them to specific multicloud requirements.
Finally, Gartner lists eight distinct categories of data integration tools, including a description and sample vendors for each one:
- Analytics Query Accelerators (AQA)
- Enterprise Service Bus
- Data Integration
- Data Virtualization
- Full Life Cycle API Management
- Integration Platform as a Service (iPaaS)
- Intelligent Business Process Management Suites (iBPMS)
- Master Data Management (MDM)
This report can help D&A leaders determine which data integration technologies to implement based on their specific multicloud needs. It’s available on the Gartner website for clients only.
To learn more about Hype Cycles and Gartner’s industry-leading research methodologies, visit the Gartner website.
Looking to see ChaosSearch in action?
Gartner’s three emerging technology trends themes
Evolving and expanding immersive experiences
The future of digital experience is immersive. A collection of emerging technologies supports such experiences through dynamic virtual representations, environments and ecosystems of customers and people, as well as new modes of user engagement.
“With these technologies, individuals can control their own identities and data and experience virtual ecosystems that can be integrated with digital currencies. These technologies help reach customers in new ways to strengthen or open new revenue streams,’’ Gartner said.
The technologies to watch that deliver evolving and expanding immersive experiences are metaverse, non-fungible tokens (NFTs), super apps and Web3, decentralized identity, digital humans, digital twin of the customer and internal talent marketplaces.
SEE: Metaverse cheat sheet: Everything you need to know (free PDF) (TechRepublic)
Accelerated AI automation
AI adoption is expanding as an integral part of products, services and solutions. This is accelerating the creation of specialized AI models that can be applied to automate model development, training, and deployment. AI automation refocuses the role of humans in AI development, resulting in more accurate predictions and decisions and faster time to expected benefits.
The technologies that are supporting accelerated AI automation are autonomic systems, causal AI, foundation models, generative design AI and machine learning code generation.
Optimized technologist delivery
Successful digital businesses are built, not bought. A set of emerging technologies focuses on the product, service and solution builder communities, such as fusion teams, and the platforms they use. These technologies provide feedback and insight that optimize and accelerate product, service and solution delivery and increase the sustainability of business operations.
The critical technologies that are optimizing technologist delivery are augmented FinOps, cloud data ecosystems, cloud sustainability, computational storage, cybersecurity mesh architecture, data observability, dynamic risk governance, industry cloud platforms, minimum viable architecture, observability-driven development, OpenTelemetry and platform engineering.
Gartner said the Hype Cycle for Emerging Technologies identifies key insights from more than 2,000 technologies and applied frameworks that the firm profiles each year into a succinct set of “must-know” emerging technologies and trends.
The report also details most of the technologies that appeared in the 2021 version of the hype cycle that the firm has removed. While they are still tracked in other hype cycles, some have been retired while others have updated profiles, Gartner said.
Subscribe to the Innovation Insider Newsletter
Catch up on the latest tech innovations that are changing the world, including IoT, 5G, the latest about phones, security, smart cities, AI, robotics, and more. Delivered Tuesdays and Fridays
Office-centric networks will give way to distributed enterprises
The first of those trends is the growth of the distributed enterprise. Driven by the massive growth in remote and hybrid working patterns, traditional office-centric organizations are evolving into geographically distributed enterprises.
“For every organization, from retail to education, their delivery model has to be reconfigured to embrace distributed services,” Groombridge said. Such operations will stress the network that supports users and consumers alike, and businesses will need to rearchitect and redesign to handle it.
In the end, Gartner expects that by 2023, 75% of organizations that exploit the benefits of distributed enterprises will realize revenue growth 25% faster than competitors.