New Measure for Network Access Speed: “FLOW”= x v^ y Mbps with z:1 overbooking.

acm-1

It gets more and more problematic to express price/ performance of mobile and fixed network access in figures that have significance, can be measured and compared in a meaningful way. It would go too far to compare this situation to the (Sjoemelsoftware) Trick Software that was discovered in the automobile industry’s Dieselgate, but blurring of facts versus PR marketing figures come very near in the Telecom/Internet billion dollar industry. What are the problems that did lead to this confusing situation?

a. Provider given parameters are vague. The concept of “bandwidth” of the access network line (twisted pair, coax cable, optical fiber pair) is outdated and less and less relevant. In its formal definition it did mean the width of the analog signals frequency band in Hertz [kHz] that could be used for voice and later modem beeps and video signals (preferably identical attenuation- flat for all frequencies and with a linear phase spectrum of the end-end signal transferred without distortion). By using the Shannon and Nyquist Laws this bandwidth could be translated into a peak datarate achievable on this link as described by ‘Kilo bits per second’ [Kb/s,  Kbps or Mbps ]. This sounds impressive and wonderful to laymen and -women, who often do not know (or do not want to know) the distinction between bits (as in line speed) and Bytes (as in computer storage and -memory), but it is rather VAGUE and meaningless as an objective specification since actual digital link speed is much lower depending on noise levels and coding and error correction employed, which come at a cost in speed.

So stating that the ADSL link has a speed of 20 Mbps should say “it has a PEAK DIGITAL CAPACITY of the downlink of UP TO 20 Mbps on copper twisted pair cables of actual maximum lenght in the ground of 400 meters from the (former) phone-exchange”, but nobody in the world bothers to  write that down for you to read. Which means that you actually get less. This is comparable with the speedometer of your car. You better not count on it that you can drive 310 km/hour with it, nor would it be possible or legal on congested roads in cities.

km-teller

And if you dare to complain that you link is slower than you expected from the ISP brochure/ site, you will be kindly referred to the small print of contracts and specifications which are open to the public to read somewhere on the planet Vogon on Friday evenings.

So actual situations/ contexts of the physical cables can vary but there is more that influences the performance and Perception of Quality of the links  you use.

b. Another fact which is hidden by the vague “marketing talk” is the fact that the Upload speed is usually much lower than the Download speed, also called ASYMMETRIC.  Well, providers when asked about that, answer that their clients do not want higher Upload speeds. Which may be true in the case of nearly all the time downloading files/ software or streaming films from central servers. But the asymmetry is intentional by the providers in order to boost download speeds trough the limited bandwidth of the copper cable or coax networks. It is however not functioning well if you use the links for business use. For instance when connecting from home with the company servers or cloud computing.

To explain that if you at home have 37 down and 14 up and the company computer has 50 down and 15 up then communications between you and the office will be 15 Mbps towards you and  14 Mbps towards the office computer. That is right , the slowest of both network access links. Which the providers have not even mentioned in your contract. Reason in simple terms: is you can write (upload) letter very slowly it is of no use for the reciever to be able to read (download) your letters very quickly. So for business use and home offices you need SYMMETRIC network access speeds. A first step towards  that is that providers be required to specify exactly what they offer, and we measure the link if the deliver with for example the OOKLA.COM  “Speedtest” App.

  • c. What is the “overbooking ratio”, which means:  how many households/ office cables are connected to one uplink further down the network? Is it 5 : 1 or 20: 1. Well that makes sense since datacom is bursty, so it is no problem that more people connect their laptop of tablet at the same time to a hub/router which combines and routes the burst of  datapackets to their respective destination.  Up to a certain point when all the kids in the neighborhood (up and) download different films at the same time and the links congest,  like roads with too many cars on them. Cable(TV) and mobile operator networks are designed with a particular contention/ overbooking ratio in mind to optimize shared use of network resources. But telco networks have the same problems now since network USE has shifted the last couple of years from telephony, fax and modem TO mobile/nomadic data, digital video streams and business computer-computer transactions (LAN-LAN & smartphone – cloud datacenters) which must have low latency and multiple Giga-bps streams with high availability. So the digital copper infrastructures we have had been implemented for some thing else which is vanishing like gramophone records CD’s and Fax machines.

I propose that Network Access providers (ISP’s, CableCo’s and Mobile Operators) are required to state in their adds what is the offered [Sustained upload and download  speeds] and actual overbooking ratio for the specific access link. In the case of the above picture it would be stated as:  Flow = 37 v ^ 14 Mbps with 12 :1 overbooking  instead of advertising it as (up to) 50 Mbps broadband. 

We should get rid of the vague marketing figures we get and get verifiable and measurable  transparency in the Network Access market. Otherwise we end up in silly Dieselgate-like smokescreens. By showing what is actually offered and delivered the providers can turn this realism into opportunities. For instance by differentiating is price/ performance for specific different client groups of businesses AND consumers, who may be willing to pay for a better Quality of Service (QoS).

 

 

 

 

 

Advertisements

About broodjejaap

See ABOUT on TheConnectivist.wordpress.com
This entry was posted in Uncategorized. Bookmark the permalink.

One Response to New Measure for Network Access Speed: “FLOW”= x v^ y Mbps with z:1 overbooking.

  1. Pingback: Van MEER naar BETERE Verbindingen | The Connectivist

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s