Artificial Intelligence in Data Centre Operations

Alibaba Cloud contributes to the open-source community

Alibaba Cloud has announced its latest contribution to the open-source community by open-sourcing its seven billion parameter Large Language Models (LLM), Qwen-7B and Qwen-7B-Chat, through its AI model community, ModelScope, and the collaborative AI platform, Hugging Face. The company has introduced its proprietary LLM, Tongyi Qianwen, earlier this year in April. This model, capable of generating human-like content in both Chinese and English, has different model sizes, including seven billion and above parameters. In an effort to democratise AI technologies, the models’ code, model weights, and documentation will be freely accessible to academics, researchers and commercial institutions worldwide. For commercial uses, the models will be free to use for companies with fewer than 100 million monthly active users. Programs with more users can request for a license.  "By open-sourcing our proprietary large language models, we aim to promote inclusive technologies and enable more developers and SMEs to reap the benefits of generative AI," says Jingren Zhou, CTO of Alibaba Cloud Intelligence. "As a determined long-term champion of open-source initiatives, we hope that this open approach can also bring collective wisdom to further help open-source communities thrive." The Qwen-7B was pre-trained on over two trillion tokens, including Chinese, English and other multilingual materials, code and mathematics, covering general and professional fields. Its context length reaches 8K. In training, this model was aligned with human instructions. Both models can be deployed on cloud and on-premises infrastructures. This enables users to fine-tune the models and build their own high quality generative models effectively and cost efficiently. The pre-trained Qwen-7B model distinguished itself in the Massive Multi-task Language Understanding (MMLU) benchmark, scoring a notable 56.7, outperforming other major models with similar scales or even some larger size models. This benchmark assesses a text model's multitask accuracy across 57 varied tasks, encompassing fields such as elementary mathematics, computer science and law. Moreover, it achieved the highest score among models with equivalent parameters in the leader board of C-Eval, a comprehensive Chinese evaluation suite for foundational models. It covers 52 subjects in four major specialities including humanities, social sciences, STEM and others. Additionally, it reached outstanding performance on benchmarks of mathematics and code generation, such as GSM8K and HumanEval.  Click here for latest data centre news.

DTX + UCX Europe 2023 prepares teams for a new realm

As 2023's advancements in technology pose more questions than answers, the industry's most insightful minds come together at Digital Transformation EXPO Europe (DTX) - the home of digital transformation, and UC EXPO (UCX) - the home of customer and workplace tech. Taking place on 4 and 5 October 2023 at ExCel London, the event comes at a time when IT teams and business leaders are starting to realise the impact from emerging technologies which has led to the theme of this year's show being ‘preparing for the next realm'. Organisations can learn how to balance risk with reward - readying their people, processes and infrastructure for the continual evolution of enterprise technology and AI. On day one of the show, three inspiring headline speakers take to the stage with Baroness Martha Lane Fox, President of the British Chamber of Commerce and Founder of lastminute.com, opening the show to explore why businesses need to decide if AI is friend or foe. Plus, world-renowned social engineer and author of ‘People Hacker', Jenny Radcliffe, speaking on the impact of tools such as ChatGPT on human-based attacks and approaches; and former distinguished engineer at Google Cloud, Kelsey Hightower, discussing the future of open source for product innovation. Day two will feature comedian and TV presenter Dara O’Briain alongside presenter & UCL professor, Hannah Fry to discuss the key role of storytelling and the value of being human in the age of machines. Furthermore, 18 stages of content will host case studies, panel debates and peer-to-peer roundtables discussing cloud, networks, cyber security, DevOps, software engineering, unified communications and collaboration, AI and automation and data, plus a new stage dedicated to IT Ops and service management. This year's event will be backed by a host of household names including Cisco, Google, Microsoft, AWS, Zoom and HP, plus Sophos, Okta, ThreatLocker and Checkpoint. Dominie Roberts, Programme Lead for DTX + UCX, says, “Following our hugely successful Manchester show in May, we're even more excited for the Europe edition of DTX+UCX returning to London ExCeL at such a crucial time for IT teams and business leaders.” "This year's event is designed to help organisations gain the knowledge and solutions they need to keep up with advancements in technology and the way a new generation of customers and employees are using it.” “Given the lineup of headline speakers, featured case studies and technology showcases, this year is sure to be one of the best yet and we’ll be announcing the details of the agenda and expo over the next few weeks.” To register, click here. Find more events happening around you here.

Fighting a surge in AI-assisted cyber attacks with the help of AI

By Richard Massey, VP of Sales, EMEA at Arcserve With cyber attacks never ending, and its impact often lethal, organisations are constantly looking for ways to enhance their data resilience. It’s a back and forth battle, good guys versus bad guys, and recently the bad guys have taken a step forward. They’re using AI to ramp up the frequency and severity of their attacks. Worse, many newbies are jumping in to try their hand at cyber crime. Script kiddies with zero coding experience can grab off-the-shelf AI tools and create and deploy malicious software. Anyone with bad intentions can quickly develop and unleash malware that wreaks havoc on companies. For instance, readily available AI tools enable even unsophisticated actors to execute denial-of-service attacks, create phishing emails and launch ransomware. These attacks can be run simultaneously from numerous systems worldwide, making it nearly impossible for human operators to manually detect all the attacking systems accessing their websites or portals. Turning AI against the hackers It’s not all bad news for the good guys. AI and deep learning technologies are also potent weapons in the fight against cyber crime. AI-driven security solutions with self-learning capabilities can proactively respond to emerging threats and protect against a wide range of attacks — effectively putting the power back in the hands of organisations. For instance, AI security tools can detect anomalies and patterns indicative of malicious behaviour and stop attacks before they cause harm. This intelligent approach to data protection reduces reliance on reactive measures and empowers organisations to stay one step ahead of cyber criminals. AI and deep learning protection systems can also adapt and evolve to counter emerging threats. They can learn from past incidents and continuously improve their defence mechanisms. By leveraging techniques like transfer learning, these systems can update their knowledge base with the latest threat intelligence and ensure greater resilience against future attacks. These systems can also take proactive, automated actions based on predefined rules or learned behaviour. For example, upon detecting a security breach or anomaly, the system can automatically trigger measures like isolating affected systems or blocking suspicious traffic. This automated response reduces the time between detection and remediation, thereby minimising the potential impact of a cyber attack. AI in action Here’s an example of what AI looks like in action. There is a well-known threat in the cyber security industry called a remote administration tool. RAT can be embedded into a simple email attachment, such as a JPEG image, allowing cyber attackers to gain unauthorised access to a system. Antivirus engines typically detect RATs based on their signatures, then distribute an alert to all endpoints to identify and remove the RATs. However, attackers can easily modify their RATs, even slightly, to generate a different signature and evade traditional signature-based detection. To fight back, AI and deep learning technologies are crucial. Instead of relying solely on static signature matching, modern cyber security tools powered by AI can analyse the behaviour of files and processes. They can observe whether a file is executing specific actions or installing software. AI security tools can flag suspicious behaviour and prevent potentially malicious actions by learning and recognising patterns in these activities. This approach is more effective in detecting and stopping emerging threats. Attackers are constantly developing new methods to evade conventional cyber security measures, which makes it essential for organisations to keep pace. AI and deep learning can play a vital role in analysing actual threats and predicting potentially malicious actions based on observed patterns. Such a proactive approach enhances the security posture of organisations and helps them protect against evolving cyber threats. A still-evolving tool When implementing AI and deep learning tools, it’s essential to consider the challenges they may bring. We’ve discussed the benefits of AI, but it’s crucial to remember that mistakes can occur. AI is still evolving and is not 100% foolproof. Sometimes, it may misinterpret what is happening, disrupting data or system availability.  These disruptions might happen when the AI detects what it thinks are illegal activities. For instance, AI tools often work with a reliability score. An organisation can take preventive actions if the score falls below a preset threshold. However, these preventive actions may be unnecessary, resulting in unplanned downtime. As an evolving technology, AI cannot guarantee absolute perfection, and the threat of errors will always exist. Nonetheless, as more people use the technology and encounter various threats, AI systems will improve and become better at distinguishing real threats from non-threatening situations. Getting Started with AI Many companies are intrigued by AI’s potential but don’t know how and where to start with the technology. The easiest way is to work with reliable security solution providers well-versed in deep learning and AI and already incorporating the technology into their existing products. This approach enables end-users to embrace AI and apply it effectively in data resilience and cyber security. As the technology continues to evolve, we expect to see more inhouse AI and deep learning solutions developed and deployed. However, AI’s complexity will take some years to become mainstream. In the meantime, the most accessible and straightforward way for organisations to use AI to defend themselves is to engage with solution providers with readily available AI-powered tools that neutralise cyber attacks and protect against data loss. Click here for latest data centre news.

Global Cyber Summit highlights Ukrainian experience amid geopolitical tensions

Nineteen Group, organiser of International Cyber Expo, has announced its programme for the annual Global Cyber Summit, sponsored by Sonatype, Opentext and Infoblox, and hosted at Olympia London on 26 and 27 September 2023. The summit returns with greater international appeal. Among other topics of discussion, guest speakers will provide the Ukrainian perspective on cyber security, in light of recent geopolitical events. With opening remarks by Professor Ciaran Martin CB, Chair of International Cyber Expo’s Advisory Council, the Global Cyber Summit assembles the industry’s great minds to review ongoing cyber threats, priorities and challenges. Uniquely, the programme this year invites advisors closely associated with Ukrainian government agencies to present their invaluable insight into the reality and impact of Russian cyber attacks on the country and beyond. Special guest speakers include, Oksana Kharchenko, a member of YouControl, who will delve into the challenges of managing sanctions risk in the current geopolitical setting; and Andrew Hural, Director, MDR of UnderDefense, who will reflect on the last 500 days of Russian cyber operations, determining the successes and failures of their espionage. Here are a few agenda highlights: Nicola Whiting MBE, co-owner of Titania Group, will reveal why diversity and inclusion efforts might be stalling and provide a new framework. Theresa Deumchen, Tech Policy Associate at Global Counsel, will examine the regulatory landscape concerning generative AI. Alexsander Gorkowienko, SecurityLabs’ Senior Managing Consultant at Spirent Communications, will explain how EU security regulations, such as the NIS 2 Directive, might affect businesses across the region. Jake Moore, Global Cyber Security Advisor at ESET, will shed light on his attempt to manipulate recruitment staff, land a job inside a company and gain full access to their data. Stewart Bertram, Head of Cyber Threat Intelligence at Elemendar, will utilise a mix of case studies and theories to expose the crossover between misinformation and cyber threat operations. Rashik Parmar, Group CEO of BCS, The Chartered Institute for IT, and Dr Saritha Arunkumar, IBM Public Cloud Worldwide Technical Leader - Security, will sit together on a panel to address the question: What does the rise of AI and quantum computing mean for the future of cyber security? Charlotte Hooper, Helpline Manager at The Cyber Helpline, will highlight the impact of cybercrime on individuals and what can be done to support them. Attendees can also take advantage of scheduled talks at the collocated International Security Expo. In fact, Joel Aleburu at Microsoft, will be speaking here about the role of cyber espionage in terrorist activities on the first day of the event, while Joe Wrieden, Intelligence Analyst at Cyjax, will assess the key role of Advanced Persistent Threats(APTs) in serious and organised crime on the second day. All sessions are CPD Certified. To register for free as a visitor: https://ice-2023.reg.buzz/dcnnAs press: https://www.internationalcyberexpo.com/press-pass-registration Click here for latest data centre news.

The Data Lab Community grows its online ecosystem

The Data Lab Community has expanded with AI and data skills in more demand than ever and interest in the sector growing internationally.   As part of its continued evolution, and to provide further value to its members, Scotland’s innovation centre for data and AI has launched a new premium membership subscription. The offering includes additional benefits without removing existing features that will remain freely available to both new and existing members.  Members who sign up for the premium tier will receive additional benefits, including discounts for The Data Lab’s flagship events, professional development opportunities and the ability to share job vacancies with thousands of students and professionals. Originally launched as a response to the quickly evolving and extremely challenging pandemic landscape, its growing community recently reached a significant milestone, surpassing 3,000 members. It has become a forum for academics, students and professionals to connect and collaborate on data and AI related projects. It has subsequently provided fertile territory for those looking to start or develop careers in this industry.   The focus on online collaboration has proven particularly attractive to the international audience, with members from 67 countries. This growth has seen the company become a central part of the innovation centre’s strategy to engage and inspire Scotland’s AI and data sector internationally.  Additionally, it has helped to facilitate more than 1,300 job opportunities for its members, the equivalent of almost 60 roles every month, since September 2021. The latest figures cover job opportunities that have been shared on a job forum to which members have exclusive access. The online community has helped to bridge the widening AI and data skills gap that threatens to restrict the growth of a burgeoning Scottish success story.  Click here for latest data centre news.

Panasas helps drive academic research initiatives

Panasas has announced that a number of leading academic research institutions are using storage solutions from the Panasas’ ActiveStor portfolio to support their modern high-performance computing (HPC) environments. Along with many university research institutions, the Minnesota Supercomputing Institute (MSI), the UC San Diego Centre for Microbiome Innovation (CMI), and LES MINES ParisTech, have trusted it to deliver the necessary data storage and management capabilities to support the HPC and AI workloads fuelling their research advancement. Modern HPC environments fast tracking academic research According to Hyperion Research, total HPC spending in 2022 reached $37bn and is projected to exceed $52bn in 2026. A key growth driver is that an increasing number of organisations and countries see the value of innovation and investing in R&D to advance society, grow revenues, reduce costs and remain competitive. High-performance data storage is the lifeblood of an academic research infrastructure. These research institutions cannot afford lost data, as it could set back the research program by months or even years. As these organisations simultaneously run and analyse data sets across multiple HPC and AI/ML applications with varied IO patterns and file sizes, they require an always-on storage solution that can manage the capacity of these bandwidth-intensive and mixed workloads while scaling to support growth. The Panasas’ ActiveStor portfolio of modern HPC storage solutions provides the reliability research teams need to keep their projects on track while eliminating the management burdens associated with roll-your-own and open-source storage systems. Its solutions currently power innovative research in agriculture, astrophysics, bioinformatics, climate research, computational chemistry, genome analysis, geophysics, high energy physics, machine learning, materials science, molecular biology and more. Here are a few examples of how Panasas is helping advance research within some of the most prestigious academic institutions across the globe: The Minnesota Supercomputing Institute (MSI) has relied on its storage for many years to support its growing HPC and AI research initiatives. Recently adding 10PB of additional storage, the institute leverages the company’s ActiveStor Ultra appliances running the Panasas PanFS parallel file system for a single-tier architecture that enables it to easily manage the unique set of mixed workloads that multiple research groups carry out every day. The UC San Diego Centre for Microbiome Innovation (CMI) has used it to build the data foundation it needed for continued research excellence, including the power to expedite data exploration and discovery with better control, usability and optimal uptime. The TU Freiberg is utilising Panasas as its parallel scratch file system. Rutherford Appleton Laboratory (RAL) team consolidated its storage onto Panasas ActiveStor and gained the scalability, performance and manageability that its workflows demanded, freeing up researchers’ time to spend time working on the data rather than managing and moving it. Today, a single IT manager spends half their time supporting 1,500 nodes. Click here for latest data centre news.

NCSC CEO warns AI must improve cyber security

The CEO of the National Cyber Security Centre has called for robust security systems in the early development of AI, amidst concerns that proper security measures are being overlooked. As businesses race to develop new AI products, a former intelligence chief explains that malicious attacks could have a “devastating” effect due to the rate AI is being developed in comparison to security. AI is set to play a huge role in many aspects of everyday life, from our homes and cities to high end national security, however, as businesses rush to develop products and secure their position in the market, the risk of misuse could act as a threat. The news follows the UK’s AI whitepaper, released earlier this year, which aims to put the UK on course to be the best place in the world to build, test and use AI tech after investing £2.5bn in the emerging technology since 2014. Suid Adeyanju, CEO of RiverSafe, comments, “AI-enabled cyber attacks present new challenges for security teams, adding increased complexities for organisations when protecting their devices and their data. Businesses evaluate their defence strategies to adapt, boosting their cyber security capabilities to ensure they are prepared for when an attack happens, either directly against their organisation or along the supply chain.” “Upskilling cyber workforces and ensuring diverse teams that can offer new and innovative ways of thinking to elevate expertise is an important part of the solution to defend against the new AI-powered threats. For businesses to protect themselves against the imminent and increasing threat of AI in cyber, strategies must be reassessed and more robust processes must be introduced.” Lindy Cameron, CEO of the NCSC, comments, "The scale and complexity of these models is such that if we don't apply the right basic principles as they are being developed in the early stages, it will be much more difficult to retrofit security." Click here for latest data centre news.

Over 1,300 business leaders declare AI as a force for good

More than 1,300 experts have signed an open letter to collectively emphasise the positive potential of artificial intelligence (AI) and relieve concerns about its impact on humanity. Coordinated by BCS, the open letter aims to challenge the pessimism surrounding AI and promote a more optimistic perspective. According to Rashik Parmar, CEO of BCS, the overwhelming support for the letter demonstrates the UK tech community's resolute belief that AI should be viewed as a "beneficial force" rather than "nightmare scenario of evil robot overlords". This comes standing in opposition to the recent letter signed by influential figures, such as Elon Musk, which called for a pause in developing powerful AI systems, citing the perceived "existential risk" posed by super-intelligent AI. The BCS signatories include experts from various businesses, academia, public institutions and think tanks, their collective expertise and insights highlight the myriad positive applications of AI. Sheila Flavell CBE, COO of FDM Group, says, “AI can play a key role in supercharging digital transformation strategies, helping organisations leverage their data to better understand their business and customers. As the UK continues to show its commitment to developing AI for good, it will help increase Britain’s position as a tech superpower and positively bolster the economy as its usage becomes widespread. In order to harness the full power of AI, the UK needs to develop a cohort of AI-skilled workers to oversee its development and deployment, so it is important for organisations to encourage new talent, such as graduates and returners, to engage in education courses in AI to lead this charge.” Hema Purohit, a specialist in digital health and social care for BCS, emphasises AI's ability to enable early detection of serious illnesses, like cardiac disease or diabetes, during eye tests. To further support Britain’s position as a global exemplar for high quality, ethical and inclusive AI practices, UK Prime Minister, Rishi Sunak will host a global summit on AI regulation this autumn. Challenges are emerging, including the potential automation of up to 300 million jobs, prompting companies to pause hiring in specific roles, but these must be approached pragmatically. Regulations will be a vital safeguard against the misuse of AI, instead of hasty and unregulated proliferation. As the world grapples with the powers of AI, these expert voices will provide valuable insights and perspectives to guide its responsible development and implementation. Click here for more latest news.

GovAssure, cyber security and NDR

By Ashley Nurcombe, Senior Systems Engineer UK&I, Corelight We live in a world of escalating digital threats to government IT systems. The public sector has recorded more global incidents and data breaches than any other over the past year, according to a recent Verizon study. That’s why it is heartening to see the launch of the new GovAssure scheme, which mandates stringent annual cyber security audits of all government departments, based on a National Cyber Security Centre (NCSC) framework. Now the hard work starts. As government IT and security leads begin to work through the strict requirements of the Cyber Assessment Framework (CAF), they will find network detection and response (NDR) increasingly critical to these compliance efforts. Why we need GovAssure GovAssure is the government's response to surging threat levels in the public sector. It is not hard to see why it is such an attractive target. Government entities hold a vast range of lucrative citizen data which could be used to carry out follow-on identity fraud. Government services are also a big target for extortionists looking to hold departments hostage with disruptive ransomware. And there's plenty of classified information in there for foreign powers to go after to gain a geopolitical advantage. Contrary to popular belief, most attacks are financially motivated (68%), rather than nation-state attempts at espionage (30%). That means external, organised crime gangs are the biggest threat to government security. However, internal actors account for nearly a third (30%) of breaches, and collaboration between external parties and government employees or partners accounts for 16% of data breaches. When the cause of insider risk is malicious intent rather than negligence, it can be challenging to spot because staff may be using legitimate access rights and going to great lengths to achieve their goals without being noticed. Phishing and social engineering are still among threat actors' most popular attack techniques. They target distracted and/or poorly trained employees to harvest government logins and/or personal information. Credentials are gathered in an estimated third of government breaches, while personal information is taken in nearly two-fifths (38%). Arguably the shift to hybrid working has created more risk here as staff admit being more distracted when working from home (WFH), and personal devices and home networks may be less well protected than their corporate counterparts. The growing cyber attack surface Several other threat vectors are frequently probed by malicious actors, including software vulnerabilities. The new Freedom of Information data reveals a worrying number of government assets are now using outdated software that vendors no longer support. Connected Internet of Things (IoT) devices are an increasingly popular target, especially those with unpatched firmware or factory default/easy to guess passwords. Such devices can be targeted to gain a foothold in government networks and/or to sabotage smart city services. Finally, the government has a significant supply chain risk management challenge. Third-party suppliers and partners are critical to efficiently delivering government services. But they also expand the attack surface and introduce additional risk, especially if third parties aren't properly and continuously vetted for security risks. Take the recent ransomware breach at Capita, an outsourcing giant with billions of pounds of government contracts. Although investigations are still ongoing, as many as 90 of the firm's clients have already reported data breaches due to the attack. What the CAF demands In this context, GovAssure is a long overdue attempt to enhance government resilience to cyber risk. In fact, Government Chief Security Officer, Vincent Devine, describes it as a "transformative change" in its approach to cyber that will deliver better visibility of the challenges, set clear expectations for departments and empower security pros to strengthen the investment case. Yet delivering assurance will not be easy. The CAF lists 14 cyber security and resilience principles, plus guidance on using and applying the principles. These range from risk and asset management to data, supply chain and system security, network resilience, security monitoring and much more. One thing becomes clear, visibility into network activity is a critical foundational capability on which to build CAF compliance programmes. How NDR can help NDR (Network Detection and Response) tools provide visibility. This kind of visibility will enable teams to map assets better, ensure the integrity of data exchanges with third parties, monitor compliance and detect threats before they have a chance to impact the organisation. Although the CAF primarily focuses on finding known threats, government IT leaders should consider going further, with NDR tooling designed to go beyond signature-based detection to spot unknown but potentially malicious behaviour.  Such tools might use machine learning algorithms to learn what regular activity looks like to better spot the signs of compromise. If they do, IT leaders should avoid purchasing black box tools that don't allow for flexible querying or provide results without showing their rationale. These tools can add opacity and assurance/compliance headaches. Open-source tools based on Zeek may offer a better and more reasonably priced alternative. Ultimately, there are plenty of challenges for departments looking to drive GovAssure programmes. Limited budgets, in-house skills, complex cyber threats, and a growing compliance burden will all take its toll. But by reaching out to private sector security experts, there is a way forward. For many, that journey will begin with NDR to safeguard sensitive information and critical infrastructure. Click here for more thought leadership.

atNorth announces highest revenue to date

atNorth has announced a group income of SEK 560 million (EUR 53 million) as it publishes its 2022 annual accounts. This figure represents a 44% increase in revenue from 2021 and further growth is expected in the coming years.  The company currently operates six data centres based in Iceland, Sweden and Finland. It is planning further state-of-the-art sites in the coming years. atNorth‘s growth calls for extensive investment and it has secured access to capital for this purpose. The investment for the year 2022 amounted to about SEK 220 million (EUR 20 million), and it is already evident that the 2023 level of investment will be significantly higher. Eyjólfur Magnús Kristinsson, CEO of atNorth, says, "In recent years, we have built an enormous wealth of knowledge. Our customers value our services and it is core to atNorth's good reputation." He also commends the support of the company‘s sponsor, Partners Group, that acquired the majority of the company's shares at the beginning of 2022. As digital transformation and the use of AI and other big data projects escalates, businesses are becoming increasingly reliant on supercomputers to perform ever more complex calculations. The resulting need for high performance data centre and supercomputing services has led to an exponential increase in demand in recent years. atNorth continues to promote the Nordic region as a superlative location for investment in digital infrastructure, recently winning the ‘Tech Capital Location’ award for Iceland as a result of its advantageous climate, favourable business environment and cutting-edge infrastructure. Eyjólfur comments, “Iceland is truly on the map, and with our best-in-class data centres, outstanding customer service and proficient marketing communications, atNorth is proud to have created a desirable position in the market.”