专家表示，人工智能的实施将重塑网络架构和程序

许多专家表示，人工智能实施的繁重网络和安全需求将需要改进架构和安全工具，而适合人工智能的工具本身可能需要由人工智能管理。

The heavy networking and security demands of AI implementation will require revamped architectures and security tools, many experts say — and tools adapted for AI may need to be themselves managed by AI.

许多专家表示，人工智能实施的繁重网络和安全需求将需要改进架构和安全工具，而适合人工智能的工具本身可能需要由人工智能管理。

"Regarding networking, there are two types of AI — AI for networking and networking for AI," industry analyst Zeus Kerravala wrote earlier this year. "The former uses AI to run the network, and the latter deploys a network to support AI."

“就网络而言，人工智能有两种类型——用于网络的人工智能和用于人工智能的网络，”行业分析师 Zeus Kerravala 今年早些时候写道。 “前者使用AI来运行网络，后者部署网络来支持AI。”

Networking and security provider Aryaka touts its Unified SASE as a Service as the optimal architecture for companies that plan extensive use of generative AI, and as Kerravala predicted, AI itself will have a role in overseeing the implementation of the service.

网络和安全提供商 Aryaka 宣称其统一 SASE 即服务是计划广泛使用生成式 AI 的公司的最佳架构，并且正如 Kerravala 预测的那样，AI 本身将在监督服务的实施方面发挥作用。

"It's the convergence of networking and security," Aryaka Senior Director of Product Marketing Klaus Schwegler told us. "A from-the-ground-up designed approach to have a unified policy, a unified way of administering, managing, orchestrating policies and have a unified control over such when it comes to networking security."

“这是网络和安全的融合，”Aryaka 产品营销高级总监 Klaus Schwegler 告诉我们。 “一种从头开始设计的方法，具有统一的策略，统一的管理、管理、编排策略的方式，并在网络安全方面对此进行统一的控制。”

Optimizing for AI …

针对人工智能进行优化...

To that end, there are two sides of the coin to Aryaka's AI strategy. One side involves tailoring networking and security to the needs of AI, using three optional features for Unified SASE as a Service.

为此，Aryaka 的人工智能策略有两个方面。一方面涉及根据 AI 的需求定制网络和安全性，使用统一 SASE 即服务的三个可选功能。

The first, AI>Perform, makes sure that network performance is optimized for AI workloads and applications. The second, AI>Secure, safeguards those AI processes by controlling access and stopping data leakage. Finally, AI>Observe gives users maximum visibility into their AI processes and network usage in general.

第一个是 AI>Perform，确保针对 AI 工作负载和应用程序优化网络性能。第二个是 AI>Secure，通过控制访问和阻止数据泄露来保护这些 AI 流程。最后，AI>Observe 为用户提供了对其 AI 流程和网络使用情况的最大可见性。

"This is crucial in an environment where real-time management and security of networks are paramount, due to the increasing prevalence of AI," wrote Aryaka Chief Product Officer Renuka Nadkarni in a recent company blog post. "The ability to observe and analyze network performance in real time allows enterprises to rapidly identify and respond to potential security threats, ensuring more resilient and robust network operations."

Aryaka 首席产品官 Renuka Nadkarni 在最近的公司博客文章中写道：“由于人工智能的日益普及，这在实时管理和网络安全至关重要的环境中至关重要。” “实时观察和分析网络性能的能力使企业能够快速识别和响应潜在的安全威胁，确保更具弹性和稳健的网络运营。”

The requirements of AI will reshape network architectures and procedures. For example, because every query to generative AI results in a unique, dynamically generated response, there’s no point in caching data. Yet because traffic to and from servers of AI content will be massive in both directions, extremely low latency is a must.

人工智能的要求将重塑网络架构和程序。例如，由于对生成人工智能的每次查询都会产生一个独特的、动态生成的响应，因此缓存数据是没有意义的。然而，由于人工智能内容服务器的双向流量都非常大，因此极低的延迟是必须的。

The solution may be to distribute AI servers geographically, somewhat like SASE points of presence, wrote Orange Group researchers Usman Javaid and Bruno Zerbib a recent piece in TM Forum.

Orange Group 研究人员 Usman Javaid 和 Bruno Zerbib 最近在 TM 论坛上发表的一篇文章中写道，解决方案可能是在地理上分布人工智能服务器，有点像 SASE 存在点。

"Future networks must expand cloud-centric architectures toward the edge, bringing LLM closer to data sources, enabling low-latency inference, improving data transfer by processing data locally, while maintaining user data privacy," they wrote.

他们写道：“未来的网络必须将以云为中心的架构扩展到边缘，使法学硕士更接近数据源，实现低延迟推理，通过本地处理数据来改进数据传输，同时维护用户数据隐私。”

… and using AI to optimize

...并使用人工智能进行优化

The other side of the coin is to use AI to extend security and networking performance in ways that human-controlled processes could not. Schwegler explained how AI could, for example, detect unusual network activity that might escape human notice.

硬币的另一面是使用人工智能以人类控制流程无法做到的方式扩展安全性和网络性能。例如，施韦格勒解释了人工智能如何检测可能逃脱人类注意的异常网络活动。

"Using AI tools in order to detect patterns, anomalies," he said. "Traffic behavior that seems abnormal. … All of a sudden, you have a traffic spike, data-transfer rates that you would not detect or too late in order to understand what exactly is happening, who is doing that."

“使用人工智能工具来检测模式和异常，”他说。 “流量行为似乎异常……突然间，流量激增，数据传输速率无法检测到，或者为时已晚，无法了解究竟发生了什么以及谁在这样做。”

Spotting such anomalies is one of several ways in which AI can boost cybersecurity, according to the U.S. Cybersecurity and Infrastructure Security Agency (CISA). Other potentially AI-assisted aspects of cybersecurity include detecting personally identifiable information (PII) and taking part in forensic examinations.

美国网络安全和基础设施安全局 (CISA) 表示，发现此类异常现象是人工智能增强网络安全的多种方式之一。网络安全的其他潜在人工智能辅助方面包括检测个人身份信息 (PII) 和参与法医检查。

"There are behavior patterns, trends that can be identified with AI much faster than any other human deterministic machine learning models that can be written," noted Nadkarni in a recent company webcast. "We also need to use AI to enhance the protection capabilities and secure access and secure all assets."

纳德卡尼在最近的公司网络广播中指出：“人工智能可以比任何其他可以编写的人类确定性机器学习模型更快地识别行为模式和趋势。” “我们还需要利用人工智能来增强保护能力和安全访问并保护所有资产。”

The role of managed service providers

托管服务提供商的作用

As a company that bridges networking and security, Aryaka sees itself as well positioned to offer its clients and users services that will permit them to maximize their AI usage safely and efficiently. The targeted market is not just direct clients but customers coming through managed-service channels as well.

作为一家连接网络和安全的公司，Aryaka 认为自己有能力为客户和用户提供服务，使他们能够安全有效地最大限度地利用人工智能。目标市场不仅是直接客户，还包括通过托管服务渠道而来的客户。

"MSPs with expertise in AI technologies will play a key role in spotting errors and ensuring the smooth integration of AI into IT workflows," Nadkarni wrote in her blog post. "AI, in turn, will play a crucial role in enhancing real-time network security by providing advanced monitoring, analysis, and error-detection capabilities."

Nadkarni 在她的博客文章中写道：“拥有人工智能技术专业知识的 MSP 将在发现错误并确保人工智能顺利集成到 IT 工作流程方面发挥关键作用。” “反过来，人工智能将通过提供先进的监控、分析和错误检测功能，在增强实时网络安全方面发挥至关重要的作用。”

Aryaka Chief Marketing Officer Ken Rutsky said in the webcast that the addition of AI assistance to networking and security tools, and the use of those tools to focus efficiency and delivery toward the best possible AI performance, will usher in a new phase of business opportunity.

Aryaka 首席营销官 Ken Rutsky 在网络广播中表示，将人工智能辅助添加到网络和安全工具中，并使用这些工具来关注效率和交付，以实现最佳的人工智能性能，将迎来一个新的商业机会阶段。

"Our goal is to help our customers get it all: performance, agility, simplicity and security without trade-offs," he said. "And as they move into these Gen AI applications, getting it all is going to become harder but even more important and more rewarding."

“我们的目标是帮助我们的客户获得这一切：性能、敏捷性、简单性和安全性，无需权衡，”他说。 “当他们进入这些新一代人工智能应用程序时，获得这一切将变得更加困难，但也更加重要和更有价值。”