Breaking Free from Legacy Observability: Why Service Providers Choose Kentik Over Deepfield
Summary
Modern network operators need modern observability tools. In this post, we explore why Deepfield — a traditional network flow analytics platform — falls short in providing comprehensive insights required for today’s network operations, and how Kentik’s modern data platform is purpose-built for today’s infrastructure teams.
Modern networks are increasingly complex, and the rate of evolution is accelerating. The move to hybrid and distributed infrastructure has made tools created only a few years ago obsolete. From carriers to hyperscalers, content providers to regional service providers, all require advanced observability tools with cutting-edge back-end data storage and query capabilities to stand a chance in today’s crowded and competitive market.
To keep up with the pace of change, business leaders and network engineers/architects need a partner who knows how modern networks operate and can provide the data platform required to ensure networks are performing well while keeping costs down. However, traditional network traffic analytics platforms with outdated architectures are not built to handle these needs. In this post, we’ll compare Kentik with Nokia Deepfield across three critical categories that define modern network observability.
About Deepfield
Deepfield, acquired by Nokia in 2017, was built to provide ISPs and carriers with network flow monitoring and DDoS defense. Deepfield correlates data sources such as IP flow records, BGP routes, and DNS info to provide visibility into on-prem traffic, services, and subscribers alongside DDoS mitigation and capacity planning. While Deepfield may meet the needs of some organizations, its limited telemetry support, legacy architecture, and opaque OTT classification are driving many to seek modern alternatives.
Telemetry support and network visibility
Modern networks are under intense pressure to push the envelope on features and performance, all while lowering costs. The only way to accomplish this is to reduce complexity and automate wherever possible. Given the increasingly complex and distributed nature of modern networks, success depends on achieving complete observability across the entire network while minimizing the number of tools required.
Kentik provides diverse telemetry support and holistic visibility across on-prem and cloud infrastructures, enabling teams to unify and correlate network traffic, performance, and device health to drive intelligent alerting, DDoS protection, and AI insights in one platform. Kentik normalizes a wide range of network telemetry sources, including IP flow, BGP routing, cloud flow logs, K8s networking, device telemetry (SNMP and streaming telemetry), synthetic testing, and a variety of business and operational metadata, to give teams full network context for faster troubleshooting and optimization.
In contrast, Deepfield lacks support for critical network telemetry types that are needed to provide this complete picture. Their platform cannot support telemetry for Kubernetes clusters or cloud-native infrastructure, such as VPC flow logs for AWS, Azure, and Google Cloud. Deepfield also doesn’t support SNMP or streaming telemetry for device/interface health, application performance monitoring, or synthetic transactions. These platform gaps create blind spots for infrastructure teams, forcing them to juggle multiple siloed tools and increasing MTTR/MTTI.
Back-end database architecture
Collecting the mountains of telemetry data is only one part of the equation. Operators need to be able to store all the data—not summaries—with fast, flexible tools to sift through it, either in an intuitive GUI or by exporting.
Kentik stores all data at full resolution with no summarization as data ages. Queries are fast and flexible, allowing engineers to find information in whatever format and level of detail they need. Kentik’s API-first architecture, seamless exporting to data lakes, custom dashboards, and flexible reporting options make it easy for any business to use AI/ML frameworks to drive data democratization.
Deepfield customers are stuck with pre-built reports and dashboards, query times that can take minutes to hours, and rigid limitations on the types of queries available. These limitations are stifling for operating a modern network and make it extremely hard—and impossible in some cases—for businesses to extract the data they need to make informed decisions. Moreover, Deepfield does not retain full-resolution NetFlow packet data, further limiting operators’ access to the information they need.
OTT classification transparency
One of the primary jobs of an observability platform is detecting and classifying various types of traffic traversing your network. With hundreds of over-the-top (OTT) services running over internet infrastructure, understanding each OTT provider’s traffic profile on your network — and digital delivery supply chain — is critical to running a modern network and key to driving multiple business outcomes.
In the past, providing more than the most basic of telemetry (e.g., “XX% of traffic is YouTube”) required Deep Packet Inspection (DPI). A few observability platforms now offer DPI-lite capabilities, reducing operational complexity and the need to run multiple appliances.
Kentik provides full transparency into OTT classification, displaying matching efficiency, certainty indicators, unclassified high-traffic hostnames, and all subscriber traffic—whether classified or not—so users can see precisely what’s being identified and why.
Deepfield’s OTT classification engine is lacking in some of these areas. Their Cloud Genome product provides some insight into OTT and CDN traffic, but the classification logic is proprietary. Users cannot know how or why traffic is classified, whether the classification is correct, or what proportion of traffic is misidentified, dropped, or inferred via statistical mapping. And if new applications come online, users have difficulty knowing how or if the application is classified.
Kentik is the modern approach to network observability
Kentik was built from the ground up by network operators for network operators. We believe infrastructure teams deserve modern observability that empowers them to optimize efficiency, reduce costs, and keep ahead of the competition. Here’s what we prioritize:
- Future-proof platform design: Kentik is built for speed and scale. Our full-resolution platform retains data at high resolution and full cardinality, delivering unbounded exploration across hundreds of dimensions OOTB.
- Diverse telemetry support: Kentik unifies on-prem and cloud traffic data, device health metrics, and synthetic performance monitoring into a single platform.
- AI-driven innovation: Kentik is pushing the boundaries of AI-powered network observability, combining deep knowledge of your network with advanced GenAI integration.
- Lightning-fast results: Kentik delivers query responses in under two seconds.
- Data, democratized: With flexible reporting and customizable dashboards, Kentik empowers everyone in your organization with real-time insights into network efficiency, performance, security, and costs.
- Complete OTT transparency: Kentik’s True Origin engine provides real-time, fully transparent classification of 1,000+ OTT services and 60+ CDNs. You’ll know exactly what proportion of your data is classified, how it’s mapped, and its accuracy – with continuous updates as internet trends evolve.
- Independent and innovation-driven: As an independent company not tied to hardware or cloud vendors, we’re committed to continuous innovation, driven by the evolving needs of our customers and the industry.
- High-touch support: Kentik’s 100% in-house customer success team provides proactive, expert support.
- Business intelligence: We enrich your network data with business context, enabling cost analytics, revenue optimization, and strategic decision-making.
Ready to explore? Book a meeting today to take the next step towards modern observability.
