Original source: NetApp
This article is an editorial summary and interpretation of that content. The ideas belong to the original authors; the selection and writing are by Streamed.News.
This video from NetApp covered a lot of ground. 6 segments stood out as worth your time. Everything below links directly to the timestamp in the original video.
Understanding the underlying network architecture is crucial for any enterprise considering hybrid cloud deployments to avoid performance roadblocks and unforeseen expenses. Learn why a pharmaceutical company's multi-cloud strategy failed due to overlooked connectivity details and how to prevent similar issues.
Hybrid Cloud Network Missteps Lead to High Latency and Unexpected Costs for Pharma Company
A pharmaceutical company performing DNA sequencing encountered severe latency issues and unexpected costs when attempting to offload on-prem data to Amazon Web Services (AWS) for processing. The project, intended to leverage AWS EC2 instances, suffered from network connectivity problems, resulting in upwards of 300 milliseconds of latency, significantly higher than the typical 2-5 milliseconds observed with cloud-adjacent storage solutions. This problem led to unbudgeted inter-regional charges as data routed inefficiently across AWS regions, highlighting a critical gap in understanding hybrid cloud networking among storage professionals and their clients.
To avoid such pitfalls, storage professionals must proactively understand cloud private connectivity and assume that customers may not have their hybrid cloud networking optimized. Leveraging dedicated interconnection platforms like Equinix provides direct, low-latency connections to various cloud providers, helping organizations maintain control over data path, performance, and costs. This ensures that data movement from on-premises environments to cloud instances like NetApp's cloud offerings operates efficiently and within budget, critical for complex workloads in regulated industries.
"It is very, very important that if you're talking hybrid multicloud or even just hybrid cloud, it's extremely important to understand the concepts underpinning how data is getting from the network interface of your on-prem to the physical and virtual network interface of the cloud."
Equinix and NetApp Partner to Deliver Composable AI Factory with Global Data Mobility and Sovereignty
Equinix and NetApp are collaborating to provide enterprises with a "composable AI factory," designed to facilitate experimentation, iteration, and deployment of artificial intelligence solutions. This integrated offering leverages NetApp's data mobility capabilities, including consistent data movement to major cloud providers via first-party offers and tools like SnapMirror. This allows data sets, such as those on an AFX cluster, to be curated through an integrated development environment (IDE) and reliably moved to public clouds, specialized NeoClouds, or edge locations.
The combined solution utilizes Equinix's global interconnection platform, Equinix Fabric, which spans 74 metros and 276 data centers, ensuring low-latency and deterministic data transfer. This setup addresses critical enterprise needs, including data sovereignty, by enabling control over network paths to keep data within specific national borders, a feature currently available in the US, Canada, and Japan. The entire infrastructure can be configured through API-driven tools like Terraform and Ansible, allowing for rapid deployment and adaptation of AI workloads, providing businesses with flexibility as their AI strategies evolve.
"Between NetApp and Equinix, customers are able to create, I'll call it, a composable AI factory, where you can pick your GPU from the platform of choice that makes sense for your economics and your performance and either move data up or use it in place."
Equinix Data Hub Eliminates Cloud Egress Fees and Vendor Lock-in for Enterprises
Equinix offers a solution for enterprises to establish a central "data hub" for their authoritative data copies, mitigating significant challenges associated with multi-cloud environments. By hosting this core data within Equinix, companies can project data to various public cloud services, such as AWS FSxN, Azure ANF, or Google Cloud Volumes, without incurring prohibitively high egress fees when transferring data between different cloud providers. This approach provides unprecedented flexibility, allowing organizations to leverage specialized services from multiple clouds and switch providers without penalty.
This strategy empowers businesses to maintain full control over their data location and movement, effectively avoiding vendor lock-in and shielding against unexpected price hikes from cloud providers. When a specific cloud service is no longer required, the associated data copy in that cloud can simply be deleted, as the primary, authoritative copy remains secure within the Equinix data hub. This enables a more agile and cost-effective cloud strategy, moving closer to a "cloud utopia" where data location and underlying infrastructure complexities are abstracted for the user.
"Equinix's value is being able to create a data hub that'll hold what I'll call the authoritative copy of that data... and moving it from cloud to cloud... no egress, no time. I'm in control."
Private Interconnectivity Critical for Agentic AI Data Sovereignty and Performance
As Agentic AI applications become more prevalent, data sovereignty emerges as a significant challenge, particularly for regulated industries. Relying on the public internet for data transfer in AI workflows means relinquishing control over data performance, path, and uptime once packets leave an organization's network, a risk deemed unacceptable for sensitive information or strict service level agreements. Traditional solutions like SD-WAN over the internet fail to provide the necessary deterministic control and security required for such critical applications.
The solution lies in establishing robust private interconnectivity. Equinix Fabric provides a dedicated, low-latency, and deterministic network that offers superior control and security, mirroring the benefits of MPLS but with greater simplicity and functionality. Equinix is developing an AI inference network with Agentic capabilities, enabling automated private connections to various AI resources, including foundation models (such as CoreWeave or Nebius), internal data centers, and SaaS providers. This allows users to define network connections through intent-based agentic workflows, abstracting complex networking tasks like BGP configuration, and ensuring data remains sovereign and performs optimally.
"Private interconnectivity really matters. And that's why a lot of companies stuck with MPLS where there was a network provider watching every link. And using Equinix Fabric, you get that same watch only with a simpler, less latent, more deterministic network from a performance perspective."
Equinix Platform Enables Flexible GPU Utilization, Eliminating Cloud Egress Barriers for AI Workloads
The economics of GPU usage in the cloud resemble a lease model, as the high cost and intermittent need for these powerful resources make outright purchase impractical for many enterprises. However, tying AI data to a single hyperscaler often creates significant egress barriers, limiting an organization's flexibility to optimize costs and performance by choosing alternative GPU providers. This challenge restricts the ability to leverage specialized "NeoClouds"—providers like Groq Cloud that offer custom chips optimized for specific AI tasks, such as inference, at potentially lower costs than traditional hyperscalers.
Equinix addresses this by hosting various NeoCloud GPU providers within its global interconnection platform. By establishing a core data presence on real storage, such as NetApp solutions, adjacent to these diverse GPU services and hyperscalers, customers can seamlessly switch between providers without incurring data egress fees. This strategy allows enterprises to match specific AI use cases with the most economically and performant GPU resources available, from Groq Cloud for inference to hyperscalers for other needs, while also connecting to legacy data centers and SaaS providers, thereby maximizing operational flexibility.
"If I put my data up there, it makes it a lot harder for me to get my data out here. It's going to cost me money to do that. So, there's a barrier to business there. I shouldn't have that barrier."
Liquid Cooling Costs and Complexity Drive Enterprises to NeoClouds and Managed Services for AI Infrastructure
The escalating demands of artificial intelligence, particularly with advanced GPU infrastructures like the Nvidia Grace Blackwell Superpod, are introducing complex and costly physical requirements such as liquid cooling. This challenge is paralleled with the high costs associated with achieving zero Recovery Time Objective (RTO) and Recovery Point Objective (RPO) in traditional storage, where enterprises often compromise on the most stringent requirements due to budget and operational realities. Many organizations, realizing the significant investment in specialized cooling and its ongoing management, are not prepared for these new operational burdens.
Consequently, enterprises are increasingly expected to leverage
Also mentioned in this video
- Equinix is a global digital infrastructure company with 276 data centers… (0:58)
- The pharmaceutical company's networking issues were resolved by re-architecting… (13:05)
- Glenn recommends that customers conduct free workshops with Equinix to analyze… (14:03)
- Equinix as a central hub for IT, comparable to Grand Central Station, where… (25:02)
- Equinix's role in the storage ecosystem, including a new tape-as-a-service… (26:17)
Summarised from NetApp · 43:03. All credit belongs to the original creators. Streamed.News summarises publicly available video content.
Streamed.News
This publication is generated automatically from YouTube.
Convert your full video library into a digital newspaper.
Get this for your newsroom →
