Pontos principais
- As estratégias tradicionais de resiliência estão se mostrando insuficientes diante da escala,da velocidade e da autonomia dos sistemas baseados em IA.
- Ransomwares industrializados e ataques baseados em IA agora têm como alvo os sistemas de backup,minando os alicerces da Recovery.
- As organizações precisam deixar de contar com equipes de segurança e Recovery isoladas e adotar um modelo unificado e contínuo,conhecido como operações de resiliência.
- A resiliência baseada em IA requer visibilidade de dados em tempo real,detecção contínua de ameaças e Recovery inteligente e limpa em escala.
- Plataformas modernas permitem uma Recovery rápida e verificada,ajudando as organizações a evitar o dilema entre restaurar rapidamente e restaurar com segurança.
À medida que as organizações correm para adotar a IA,os CISOs e CIOs estão chegando a uma constatação gritante: as estratégias de resiliência que funcionavam para a infraestrutura tradicional estão falhando. Sistemas que conseguiam se recuperar de ataques em horas agora podem levar dias. Abordagens de backup projetadas para dados centralizados podem ter dificuldades com cargas de trabalho distribuídas entre nuvens e plataformas de IA. Enquanto isso,as ameaças e as vulnerabilidades potenciais crescem a cada dia.Em um webinar recente,Tim Zonca,vice-presidente de marketing de portfólio da Commvault,abordou uma questão urgente enfrentada pelos líderes de segurança: como manter a resiliência quando a IA muda fundamentalmente as regras?
What Industrialized Ransomware and AI Mean for Resilience
CISOs and CIOs are under pressure. In spite of billions spent on cyber defense,nation-states and professional crime rings continue to reap ever larger payoffs from their victims. Ransomware-as-a-service has become widespread,and advanced AI automation is accelerating the industrialization of malware. By including backup systems in their attacks,adversaries are undermining the very foundation of resilience.
As attacks become more sophisticated,targets are becoming more vulnerable. AI is scaling faster than organizations can secure,with exponential data growth,fragmentation across environments,more complex supply chains,and autonomous systems operating with minimal oversight. AI agents and non-human identities now outnumber humans 80 to 1. When these systems make mistakes or expose vulnerabilities,the impact can cascade across interconnected business processes.
Breaches and failures are now almost inevitable; the only question is whether you can recover fast enough to keep your business running. For organizations using legacy systems that assume human-controlled systems,centralized data,and isolated failures,the answer may well be no.
Making Resilience Operational
As AI agents make decisions across the environment,including a significant number of errors,it’s no longer enough to focus on protecting infrastructure. Security leaders must now broaden their operational focus across three critical areas:
- Proteger continuamente os dados na fonte e monitorar anomalias.
- Controlar as identidades de pessoas,identidades não humanas e dispositivos que acessam e utilizam dados de forma autônoma.
- Alcançar a recuperação previsível de dados em grande escala,sem comprometimento ou corrupção.
Tradicionalmente,a segurança de dados,a resiliência de identidade e a recuperação cibernética têm funcionado como disciplinas independentes,cada uma com sua própria equipe,ferramentas,políticas e requisitos. Esses silos deixam vulnerabilidades para os invasores explorarem e retardam a recuperação quando os sistemas de IA falham. Para preencher essas lacunas,as organizações devem unificar esses recursos em um ciclo contínuo e automatizado. Chamamos essa abordagem deoperações de resiliência (ResOps).
O ResOps abrange três requisitos essenciais para a resiliência da IA:
- Understanding your data landscape: Knowing where data lives,its sensitivity,who’s accessing it (including AI agents and non-human identities),and what policies govern that access in real time. For AI workloads,this extends to protections like LLM prompt governance to control how models access data.
- Continuous threat detection: Automated systems that constantly monitor for anomalies,compromised identities,and data corruption. When AI systems are making thousands of autonomous decisions,you can’t wait for periodic security reviews.
- Intelligent recovery: Automated,comprehensive restoration for entire cloud-native applications and their dependencies. To prevent re-infection,teams must validate data integrity and conduct forensic analysis in an isolated cleanroom before moving trusted data back to production.
Enabling ResOps in practice
To help companies make the move to ResOps,Commvault has introducedCommvault Cloud Unity,o lançamento de plataforma mais significativo de nossa história. Ele foi projetado para reunir as três dimensões da resiliência:
- Reunindosegurança de dados,resiliência de identidade,and recuperação cibernéticasob um único modelo operacional.
- Protecting all workloads,from both today’s production systems to tomorrow’s emerging AI stacks.
- Safeguarding data regardless of location,whether in clouds,regions,data centers,or edge locations.
A next-generation architecture brings AI automation to all facets of data protection,segurança de dados,resiliência de identidade,and recovery. For security and IT teams,the platform provides simplicity at scale with one experience,one policy engine,and one interface designed to protect data,predict threats,and accelerate clean recoveries.
As security leaders know all too well,recovering from the most recent backup minimizes data loss but risks restoring compromised data. Rolling back to a verified clean state may eliminate threats but means losing hours or days of business-critical transactions or AI model training.
With Commvault Cloud,o monitoramento contínuo de ameaças and verified clean recovery points help eliminate this forced choice. The platform architecture automatically maps dependencies across distributed systems,helps maintain immutable backups,and helps enable one-click restoration of entire environments. Recovery can be both fast and clean,helping minimize loss as well as risk.
See ResOps in action
Assista ao webinar completo sob demanda to learn more about ResOps,explore the architecture and services of Commvault Cloud,and rethink your resilience strategy for the AI age.
FAQs
Q: What is Resilience Operations (Res Ops)?
A: ResOps is an operating model that unifies segurança de dados,resiliência de identidade,and recuperação cibernética into a continuous,automated discipline rather than treating them as separate IT functions. ResOps transforms resilience from a reactive response to incidents into an active practice that continuously understands data access patterns,helps detect threats and anomalies,and enables fast,intelligent recovery at scale.
Q: Why can’t traditional backup and recovery handle AI workloads?
A: Traditional backup tools were designed for centralized,human-controlled systems with isolated failures. AI workloads involve autonomous agents accessing distributed data across clouds and complex dependencies between microservices and containers,and they operate at a scale that manual processes can’t match.
When AI systems fail or are attacked,you need to recover not just data but entire application infrastructures with all their configurations,policies,and relationships – capabilities traditional backup tools lack.
Q: What does “unified resilience” mean in practice?
A: Unified resilience means bringing segurança de dados,identity management,and recuperação cibernética together under a single platform,policy engine,and operational model rather than managing them as separate functions with different teams and tools.
In practice,this provides a consistent approach to protect all workloads and data locations,automatically correlate security events with access patterns,and orchestrate comprehensive recovery that restores both data and the complete application infrastructure needed to use it.
Q: What’s the difference between cyber resilience and AI resilience?
A: Cyber resilience focuses on protecting infrastructure and recovering from security incidents,treating resilience as an operational state for confronting threats. AI resilience expands this to address challenges unique to AI-driven systems: autonomous agents making decisions with minimal oversight,exponential growth of data and non-human identities across environments,and cascading failures where problems in interconnected AI systems impact entire business operations rather than staying isolated.
Q: How does ransomware target backup systems?
A: Ransomware increasingly targets backup systems by exploiting compromised credentials with privileged access,moving laterally from production systems to connected backup repositories,or exploiting vulnerabilities in backup software itself. Modern ransomware families specifically hunt for backup infrastructure to encrypt or delete recovery points,preventing organizations from restoring clean data and maximizing pressure to pay ransom. This makes offline,immutable,or air-gapped backups essential for resilience.
Q: What is the clean vs. complete recovery dilemma?
A: The clean vs. complete recovery dilemma is the forced choice organizations face during incident response. You can recover from the most recent backup to minimize data loss but risk restoring compromised or corrupted data; or you can roll back to a verified clean state before the incident to eliminate threats but lose significant business-critical data. Traditional backup tools make organizations choose between completeness and safety,while modern resilience platforms aim to provide both simultaneously through o monitoramento contínuo de ameaças and verified recovery points.
Q: What is a cleanroom in recuperação cibernética?
A: A cleanroom in recuperação cibernética is an isolated,secure environment completely separated from production systems to help organizations safely test,validate,and analyze recovered data before restoring it to active use. Cleanrooms help enable forensic investigation of compromised systems,testing of recovery procedures,and verification that restored data is free from malware or corruption – all without risking reinfection of production environments or exposing sensitive data during analysis.
Sam Curcuruto is Director of Product Marketing at Commvault.
Blogs relacionadosRepensando a resiliência para a era da IA
A CIO’s Perspective: Strengthening Business Resilience in the AI Era
Resiliência contra a máquina de IA
Cleanroom Recovery abrem caminho para uma nova era em resiliência cibernética
