Riverbed was initially best known for and defined by its Steelhead product line for network and application acceleration. Over the last several years, the company has embarked on an aggressive set of updates designed to help improve not only network acceleration but also network observability. At the core of those updates is a set of AI-powered capabilities that benefits from the Riverbed Data Store, which is the company’s proprietary data layer.
Today, Riverbed announced new AI-driven intelligent network observability solutions featuring next-generation xx90 hardware appliances, flexible subscription licensing, and an integrated AI bundle called Network Observability Essentials. The new updates follow a series of releases from Riverbed in 2025, including AIOps and network acceleration technologies.
The announcement comes as Riverbed reports 92% year-over-year bookings growth in observability during the first half of 2025. This latest release represents more than new products; it marks the completion of a platform transformation that has unfolded over the past two years under CEO Dave Donatelli‘s leadership.
“In the two years since I’ve been here, we’ve pretty much revitalized and brought out new products across the portfolio,” Donatelli told Network World. “So we are innovating at a very rapid rate.”
From WAN optimization pioneer to AI-powered observability platform
When Donatelli joined Riverbed in July 2023, the company possessed powerful but underutilized technology assets. Known primarily for its Steelhead WAN optimization appliances, Riverbed had developed sophisticated data collection and analysis capabilities over more than two decades, but hadn’t fully recognized their potential in the emerging AI-driven observability market.
Donatelli recounted that while he was doing his due diligence on Riverbed before joining the company, he learned that it had figured out a unique way to handle very large scale data. That technology is the Riverbed Data Store, and in the modern AI era, it’s even more valuable.
“So, in the world of AI, as you know, data is everything, right?” Donatelli said. “So good data makes good AI, bad data makes bad AI.”
Data is at the foundation of Riverbed’s product updates, with its April release of the Aternity Digital Experience Management (DEM) technology, which is all about optimizing endpoint application delivery, for example. Data also helped to inform the company’s May update of its traditional network acceleration platform.
“Acceleration today has taken on a new meaning in the world of AI, because people now are moving huge data sets that are overwhelming their networks,” Donatelli said. “Our acceleration product can cut that time down by huge percentages, right? We’re saving people days, not just hours.”
What the Riverbed Data Store is all about
Central to Riverbed’s platform evolution is the company’s approach to handling observability data at scale. Rather than building a centralized data lake that aggregates petabytes of information, Riverbed developed a proprietary architecture called the Riverbed Data Store that functions more like an intelligent indexing system.
“We started thinking about this data store as a kind of index of indexes,” Riverbed CTO Richard Tworek explained to Network World. So, it’s the ability for us to know where the data is without actually moving the data.”
This architecture enables the company to move AI to the data rather than the traditional approach of moving data to centralized AI systems. The data store maintains minimal metadata, while AI algorithms can dynamically retrieve detailed information on demand. The approach addresses both performance and security concerns that plague traditional observability platforms.
“It’s more secure, because it’s a closed loop system within the customer’s environment,” Donatelli noted. “It’s only their data. We’re not intermingling their data with anybody else’s data.”
AI integration and automation capabilities
Riverbed’s AI implementation addresses network management challenges under the Riverbed IQ platform. The platform incorporates four distinct AI approaches:
- Predictive AI: Analyzes historical patterns to forecast network issues before they impact users.
- Causal AI: Determines root cause relationships between network events.
- Generative AI: Creates automated responses and documentation for common issues.
- Agentic AI: Functions as a library of skills that can be triggered by a service to execute a function.
The AI capabilities extend beyond detection to automated resolution.
“We have people deflecting, you know, hundreds of thousand of calls per month through our automation technology,” Donatelli explained. “We have a customer, for instance, who’s in the middle of a Windows upgrade, and they got flooded with errors, so we automated it to a one button resolution.”
Riverbed platform’s latest evolution
Donatelli outlined four key components in the August announcement, each addressing specific customer demands for better performance, deployment flexibility and operational simplification.
New xx90 appliance series
On the hardware side are the xx90 appliances for the company’s AppResponse, NetProfiler and Flow Gateway network observability services. Features in the updated appliances include:
- AppResponse deployments: Sustained packet capture over 50 Gbps with concurrent analysis capabilities.
- Scalable storage: Modular configurations exceeding 2.4 PB(petabytes).
- Enhanced software: AppResponse 11.21 adds real-time analysis of encrypted IPSec ESP tunnel traffic and cipher hygiene monitoring.
- Flow monitoring: NetProfiler and Flow Gateway deliver 3x faster reporting and analysis with 2x faster throughput.
Network Observability Essentials bundle
Riverbed is now also bundling its network observability software. The package includes:
- Riverbed IQ: SaaS-based AI algorithms for automated issue identification and remediation.
- Workspaces: Role-based dashboards unifying packet, flow, and endpoint data.
- Grafana Plug-In: Native integration for visualizing Riverbed data in existing Grafana deployments.
- Topology Viewer: Dynamic network mapping correlated with application and user performance metrics.
Flex subscription licensing
Lastly, Riverbed is also restructuring how organizations can choose to license its technology.
“Our license for the software, which is really the whole value of the product or operating system, is now a subscription license that you can run anywhere you want,” Donatelli said. “So, you can run on top of the hardware. You can run it virtually, or you can run it in the cloud.”
Source:: Network World