AI Race: AWS Has Created an AI Platform Offering Many Features

Telemetry data is more finely granular and frequent in the optical access network and so can help predict network problems and take preventive action without compromising the OLT’s performance, giving operators greater visibility and insights into the network. Nokia has teamed up with Rohde & Schwarz, the testing and measurement specialist, to explore the possibility of adding software-based, network measurement tools to its drone platform. Nokia already uses drones to measure networks’ performance, by attaching a smartphone loaded with Rhode & Schwarz’s QualiPoc 4.9G/LTE and 5G network measurement capabilities to the drone. That architecture exposes telemetry and control interfaces with predefined service models (known as the E2 interface) for Open RAN vendors to implement in support of near-real-time xApp applications which help operators monitor Open RANs. He continued, “We will measure our progress in this transformation based on the standards set by TM Forum for the level of automation in network operations. Our ambition is to move from level one or two currently, depending on the country, to level four by 2025, specifically on two key processes, the test integration deployment process, and the network monitoring and fault management process.

“Ideally, you will have a minimum viable product (MVP) already developed and your eyes set on seed stage funding, with plans to raise in the next 18 months. This comes as the world is anticipating more disruptive changes in the business landscape due to generative AI on the back of its ability to analyse vast amounts of data and generate new ideas. OpenAI’s launch of advanced AI chatbot ChatGPT may have represented ground zero for generative AI hype, but the company had actually formed its own early-stage investment fund as far back as mid-2021, in partnership with its largest investor, Microsoft. The announcement from AMD follows the global value of AI investment falling dramatically in 2022. Despite this, AMD’s shares fell after its company presentation event as it didn’t disclose a full customer list for its new chips. AMD revealed its new chip, the MI300X, at a presentation event this week, detailing some technical specifications that beat its Nvidia rival.

Many practitioners (Python process) tend to run inference sequentially when the server is invoked with multiple independent requests. Although it’s easier to set up, it’s usually not the best practice to utilize the accelerator’s compute power. To address this, DJLServing offers the built-in optimizations of dynamic batching to combine genrative ai these independent inference requests on the server side to form a larger batch dynamically to increase throughput. All the requests reach the dynamic batcher first before entering the actual job queues to wait for inference. You can set your preferred batch sizes for dynamic batching using the batch_size settings in serving.properties.

Experience Omniverse in the Cloud

This is useful in things like city planning – for example, seeing the impact of a new road layout or what would happen if a river bursts. This update has been much needed, making A/B testing possible within SageMaker Endpoints! SageMaker Endpoints has supported deploying multiple variants of the same model for at least the past 4 years, as well as allowing you to have a version of the model trained on different data or two different types of models deployed.

The designs are brave and bold, not necessarily realistic just yet, but sometimes that’s all we need to trigger conversations with clients and push the creative boundaries. However, you had to split the amount of traffic to each model by percentage, which made comparison of individual predictions difficult since the models might not be deterministic. With this new capability, you can send the same traffic to all your models, allowing you to compare the results more accurately.

Is AI about to revolutionise software development?

The rise of a new generation of AI-chatbots will generate a lot more messaging traffic that will be carried on telcos’ networks, as text messages are made audible and the communication becomes conversational. There are also big implications for customer service if finally chatbots can take automated customer service to a whole different level, including personalisation. Despite these travails, the total market for network automation was $3.62 billion in 2021 and is expected to grow at 24.9 percent CAGR to reach $21.45 billion by 2029, according to Maximize Market Research – see graphic below for highlights. According to 1&1, the delay is nearly all Vodafone’s fault, rather than anything to do with tech teething troubles. 1&1 is leasing tower space from several providers, but mostly from Vantage Towers whose majority shareholder is Vodafone – and operator Vodafone Germany and 1&1 are rivals.

• It delivers three times higher compute performance, up to four times higher throughput, and up to 10 times lower latency compared to first-generation AWS Inferentia.
• As shown in the following diagram, with the tensor parallel degree set to 4, the LMI container launches three Python process groups, each holding the full copy of the model.
• VMware launch its Telco Cloud Service Assurance 2.0, which DISH Wireless will test to monitor its distributed-cloud infrastructure and collect metrics and logs for actionable insights and to facilitate closed-loop automated provisioning and management.
• Currently, it is home to 13 AI companies and plays a crucial role in the regional applied AI community.
• AWS accelerator partners will also deliver dedicated content as part of the AWS Startup Loft Accelerator.
• Last year, the company was designated a European unicorn after closing a Series C investment round of USD 120 million, with a valuation of USD 1 billion.

Tilo Splitt, Head of Radio Network at Drei Austria, said, “We needed a solution that could span all our networks…Already, we’ve seen an improvement in KPIs”. In Sophia Antipolis, an area in the south of France known for considerably better weather, ETSI group on Experiential Networked Intelligence (ISG ENI) published a White Paper describing the design of a novel cognitive network. It explains how the novel system architecture (based on ETSI GS ENI-005) intelligently manages, predicts, adjusts and optimises network behaviour using cognition management to improve the operators’ experience.

Founder of the DevEducation project

I also liked how the project acknowledges other similar related open source projects (some of which I have featured in this newsletter). We publish industry press releases as a service to the tech community – please note, publication does not act as an endorsement. IBC sits at the global crossroads of the media, entertainment and technology industries providing an informative, innovative and engaging experience.

To end where we began, with ETSI, which on 28 November launched a telemetry standard designed to improve automation and deliver better customer experience of optical networks. The standards body claims it provides better monitoring than established methods via automated real-time data collection, providing greater speed and scale. Microsoft Azure will continue as OpenAI’s exclusive cloud provider genrative ai for ChatGPT as OpenAI uses Azure to train all its models. Further, Microsoft will deploy OpenAI’s models across all its consumer and enterprise products and introduce “new categories of digital experience” built on the tech. Ml.trn1 instances are powered by the Trainium accelerator, which is purpose built mainly for high-performance deep learning training of generative AI models, including LLMs.

The AWS Global Fintech Accelerator, created in partnership with NVIDIA, is dedicated to helping jump start early-stage fintech startups that are leveraging AI/ML to innovate financial services. In addition, the selected startups will receive up to $25,000 in AWS Activate credits to build their products and services. With Generative AI tools like AWS CodeWhisperer helping developers create code faster, it is nice to see that data scientists can now use similar tools to help them in their Juypter notebooks. Jupyter AI is an official subproject of Project Jupyter, that provides a user-friendly and powerful way to explore generative AI models in notebooks and improve your productivity in JupyterLab and the Jupyter Notebook. It features a native chat UI in JupyterLab that enables you to work with generative AI as a conversational assistant. It support for a wide range of generative model providers and models (AI21, Anthropic, Cohere, Hugging Face, OpenAI, SageMaker, and more).

All of which meant it was unable to install services, process payments or provide customer care. Wallaroo.AI and VMware have a strategic agreement to provide a unified edge machine learning (ML) deployment and operations platform. They say it is the first such platform that is tailored to the needs of global communications service providers (CSPs). Likewise, while previously the on-premises segment accounted for 55% of network automation deployments, cloud-based network automation solutions are rapidly gaining popularity and are forecast to grow at 25% over the next ten years.

Solutions

There are no upfront commitments or fees the Builder, and customers pay only for the AWS services used to manage their network. Amdocs, Cloudify, Infosys, Mavenir and O2 Telefónica are all investigating the Builder’s possibilities. Still across the Atlantic, Dish Network, the new national entrant with a 5G network is taking a similarly innovative approach to Rakuten Mobile and 1&1 for its network. It has strong ties with Rakuten Symphony but is under the cosh too after a “cyber-security incident” that began in February. This caused a weeks-long outage that hit the company’s internal servers, corporate communications, and some Dish-owned websites and IT telephony systems.

Collaboration software provider Mural is one of the few Gradient portfolio companies to disclose a unicorn valuation, though this may well change in the coming months when investment ramps up in the space. Based in Japan, he joined Annapurna Labs even before the acquisition by AWS and has consistently helped customers with Annapurna Labs technology. His recent focus is on Machine Learning solutions based on purpose-built silicon, AWS Inferentia and Trainium. Large language models with billions of parameters are often too big to fit on a single accelerator.

Her roles have allowed her to not only harness the power of cloud solutions but also to lead FSI customers toward innovation.As a thought leader, Victory dedicates her time to sharing her insights and expertise. Victory is a thought leader in the data space and has written whitepapers and blogs on data mesh as well as multiple content and speaking engagements relating to event-driven architectures and data democratization. Victory’s interests outside tech lies in mental health & wellbeing, youth empowerment and social action.