Most of our listeners are probably familiar with network interface cards (NICs, if you're savvy). But what about SmartNICs?
Fiserv's Michael Wynston is back with a full explainer on what SmartNICs are and how they differ from your run-of-the-mill network interface card.
Greg and Michael look closer at this emerging technology—and review why WAN managers and those who serve them might want to take note.
Subscribe to access all of our episodes:
Apple | Google | Spotify | Stitcher | TuneIn | Podbean | RSS
Key Takeaways
SmartNICs are designed to offload CPU-intensive network functions.
The development of SmartNICs was driven by the limitations of general-purpose CPUs in processing the high number of packets per second required for Network Function Virtualization (NFV).
By moving these tasks to dedicated processors on the network interface card, SmartNICs allow the main CPU to handle other processes and leverage more network interface card throughput.
SmartNICs can enhance existing NFV capabilities and troubleshooting.
While they perform functions that aren't necessarily new, SmartNICs increase the efficiency and ease of these tasks.
For example, a SmartNIC can offload generic networking functions, allowing a specialized VNF (like an SD-WAN) to focus on its unique features such as QoS or application qualification.
Additionally, many SmartNICs come with built-in capabilities for packet tapping or spanning directly off the network interface card, making it easier to perform real-time deep packet inspection, troubleshooting, analytics, and monitoring within virtualized environments. Some SmartNIC technologies can even support running a full network operating system directly on the card, bringing networking intelligence closer to the host.
Adopting SmartNICs presents challenges, particularly in brownfield environments.
One major hurdle is the increased management complexity, as distributing functions across multiple compute nodes with SmartNICs exponentially increases the number of management points compared to managing a single traditional device.
Organizations must figure out new ways to manage security policies and monitoring. Cost is another factor, as SmartNICs are not cheap. Integrating them often requires newer server hardware to take full advantage of their speed, meaning widespread adoption might involve replacing servers rather than just swapping cards.
Despite these challenges, SmartNICs offer potential benefits like value engineering in data centers by moving intelligence to the edge and allowing for simpler, less feature-rich (and potentially lower-cost) switches in the middle. For service providers, they offer the potential to deliver more services faster and cheaper at scale by running functions in hardware. Adoption in enterprise is expected to be gradual (potentially 12+ months out) due to the need for testing to validate stability, feature parity, performance, and overall value proposition before production deployment.
Greg Bryan
Greg is Senior Manager, Enterprise Research at TeleGeography. He's spent the last decade and a half at TeleGeography developing many of our pricing products and reports about enterprise networks. He is a frequent speaker at conferences about corporate wide area networks and enterprise telecom services. He also hosts our podcast, TeleGeography Explains the Internet.