Application aware latency management
To understand application performance, we must understand how latency changes over time.
Manage, analyse, and improve application outcomes with Domos technology
Domos makes software for Internet Service Providers (ISPs) that can be embedded into any customer premises equipment (CPE) for improved network quality by monitoring, analysing, and improving application outcomein real time.
With the network x-ray agent, ISPs gain access to valuable insights into any bottlenecks in their subscribers' WiFi or broadband connections. ISPs can enhance the customer experience by allowing subscribers to use this information for improving their internet through network optimisation preferences using integrated applications.
How it works
- Single Slim CPE-installed Software Module
- Latency Monitoring
- Latency Analytics
- Latency Root Cause Analysis
- Quality of Outcome (QoO) Overview & Calculation
- Context Aware Latency Management
- QoS Class Detection
- Pattern Recognition
1. Single slim CPE-installed software module
Domos’ technology provides a low-friction entry point to a spectrum of latency management capabilities all deployed upon the standards-based open customer premises equipment (CPE).
- All the magic on the CPE
- No Cloud Cost
- Domos CPE SDK is easily integrated into OpenWrt/prpl, RDK-B, OpenSync
- Supports top chipset vendors and collaboration with many OEMs
- Domos CPE API is easily integrated with TR-069, USP and similar allowing the ISP to leverage the insight Domos continuously generates
2. Latency monitoring
Domos combines a series of latency monitoring techniques with the BBF TR-452 mathematical framework for network quality such as:
- Active probing with TWAMP Light / STAMP / Ping (2)
- Dublin Traceroute
- TCP SYN ACK / DNS Lookup RTT Capture
- Real-Time WiFi Latency Distribution Estimation based on the WiFi AP based on radio environment data
The QED framework allows Domos to combine the latency sources while adjusting for noise and differences in measurement techniques. As well as monitoring segment by segment (in-home, access, core, to servers) and add them together in a mathematically sound fashion for end-to-end analysis.
Latency monitoring will be a key part of customer support, self-help, network management and capital expenditures (CapEx) analysis.
Latency analytics
Discover end-to-end root causes.
4. Latency root cause analysis
Find what is causing detrimental latency with continuous monitoring. Domos is able to pinpoint the causes of detrimental latency. What latency is detrimental is more challenging than one might think, as stable high latency may be better than lower latency that varies a lot.
From the customer premise, Domos has visibility of both in-home and the internet access and automatically separates variable and fixed latency at each segment.
Latency Root Cause Analysis will be a key part of Customer Support, Self-Help, Network Management and CAPEX Analysis.
5. Quality of Outcome (QoO) overview and calculation
Combining sources of latency data. Adding them together in a mathematically sound fashion and applying the TR-452 standard and Domos (to be standardised) distance measure, allows us to go from highly complex latency statistics to understandable statements about network quality.
Domos calculates the QoO for real traffic and what the quality would have been if a user were to initiated an application continuously. Letting ISPs truly understand the quality they deliver to subscribers.
Context aware latency management
Discover end-to-end root causes.
7. QoS class detection
Domos leverages pattern recognition to, in near-real time and within the customer premise, classify video conferencing, online/cloud gaming, video streaming, spatial computing, and AR/VR. No DPI needed or significant compute/memory resources needed. Runs easily on modern customer premise equipment (CPE) and classifies with high precision.
The algorithm observes metadata such as packet sizes, frequencies, volumes, protocols, ports, and timing. No snooping or IP based classifications which simplifies privacy, security, and is future-proof with regards to the ever increasing share of traffic that is encrypted.
Each class is based on traffic patterns and latency requirements, not content provider, hence it is compatible with reasonable traffic management under net neutrality regulations.
8. Pattern recognition
