Most of the articles I write are about general communications industry issues. For a change, I have put together a short description of the services that my company and my colleagues at Predictable Network Solutions Ltd (PNSol) jointly offer.
What we do is unique. We can:
- relate network user QoE to the supply chain behind it (through high-fidelity measurements), and
- allow operators to take control over QoE, cost and risk (through rigorous modelling and new mechanisms), and
- do so based on the reasoning and abstraction power of a robust underlying mathematical science.
There are other people who do parts of the puzzle, but nobody else does it all! These techniques have been repeatedly proven over many clients across a period of over a decade.
Selected clients
These results were all achieved using a combination of the “network x-ray” technology and the ΔQ calculus for performance modelling.
BT
We optimised the 20CN network planning rules and found 30% capex saving. (How? ΔQ measures gave visibility of true QoE and how “hot” the network could safely be run.)
Tier 1 MNO
We moved performance from 2nd/3rd in market to 1st in two weeks. (How? ΔQ measures gave visibility of a packet fragmentation issue invisible using standard low-fidelity approaches.)
Tier 1 MNO
We isolated the root cause of failure of a product line; this rescued value of £50-250m. We identified a >40% performance improvement. (How? ΔQ measures combined with the scientific method isolated the performance problems and their root cause.)
Kent Public Service Network
Optimised network planning rules and got 1 year capex deferral. (How? ΔQ measures! Some example data is here.)
Tier 1 MNO
Resolved a pre-litigation contract dispute. Estimated value of over £10m. (How? ΔQ measurements gave both the spatial and temporal resolution to be able to “see” the problem.)
Inter-telco managed service contract
Identified cause of service issue affecting ~1m customers and bad regulatory performance evaluation. (How? ΔQ measures provided visibility of a previously unknown scaling issue with TCP/IP.)
CERN
We gained a 10x improvement in data throughput from the Atlas particle supercollider experiment. (How? ΔQ measures revealed how one supposed optimisation was having the opposite of the intended effect – see this PhD thesis.)
Boeing/US DoD
We unambiguously demonstrated Future Combat Systems programme was infeasible. Saved US taxpayer >$250bn. (How? The story is here. It’s where ∆Q was invented.)
Advanced network R&D
We are also involved in several advanced R&D programmes. PNSol recently joined the EU-funded PRISTINE consortium, investigating RINA as a successor to TCP/IP. They are also funded by two current Innovate UK grants to explore new network measurement techniques. I am on the advisory board of two EU-funded research projects (reTHINK and ARCFIRE). We also do work in the field of government communications.
Benefits and costs of change
Network operators typically have an “easy” 1 year of growth “slack” in their planning rules. We can locate this capex saving. Optimal scheduling extends this to 3+ years of “free” growth (assuming a typical traffic mix). “Free” growth means no truck rolls, which is where much of the big cost savings lie. We have the tools & processes to achieve this.
Large sums are being spent on infeasible or mis-engineered systems (e.g. small cells, RAN sharing, VoIP & UC). We can identify these costs and avoid these failures.
Operators are failing to manage supply chain performance, leaving them with residual integration risks and costs. We can eliminate these by re-contracting and policing performance.
The costs of our solution include: probes, management system, licensing advanced scheduling algorithms, training & organisational development, OSS/BSS integration of more advanced performance management systems.
How to use our technology
- M&A due diligence: Know exactly what you are buying with M&A, and how much “slack” (if any) exists in the assets, and how much “hotter” they can be safely run. Understand the undocumented performance hazards in the existing products.
- Internal cost optimisation: Use fewer resource inputs into the network, and safety substitute cheaper suppliers or services whilst maintaining service quality. Combine products & rationalise the portfolio by constructing “performance safe” product substitutions.
- Steer the business using better KPIs: Run the business in a demand-led manner, focusing on the quality of the delivered experience, not the quantity of technical inputs.
- Avoid contract disputes: Know exactly where real-time performance problems are occurring in the supply chain, and have proof that would stand up in court.
- Avoid infeasible projects: Model performance hazards and treat them as integral to design, not as an after-thought. This bring us to the level normal in other engineering disciplines.
For the latest fresh thinking on telecommunications, please sign up for the free Geddes newsletter.
Read It Now
