Average per-bit delivery cost

Average Per-Bit Delivery Cost, or APBDC, is the cost accounting method by which Internet Service Providers (ISPs) calculate their cost of goods sold. [1]

Concept edit

 
Example of calculation of Average Per-Bit Delivery Cost for a small data-flow.

Average Per-Bit Delivery Cost divides the cost of however many bits were actually modulated across a network or component of a network over a period of time by the total operational and amortized capital expenses associated with the network or component over the same time-period, to produce an average cost for the delivery of each transmitted bit of data.[2] This is contrasted principally with flat-rate and 95th percentile functions, which are typically used for billing customers rather than cost accounting. In typical use, APBDCs of individual links or components of an operating network are compared with the APBDC of the whole, in order to evaluate the efficiency of the components and to track improvements in efficiency.[3][4]

High APBDC may reflect either a high cost or a low utilization, either of which is detrimental to the price/performance of the network as a whole, and indicates an area that needs attention and improvement. Besides reduction in costs of existing network components, use of already-amortized or less expensive components, and transmission of greater volumes of traffic, fundamental or "revolutionary" changes to the topology of a network, such as the substitution of peering-derived bandwidth for purchased transit, often have substantial impacts on a network's APBDC.[5]

The use of Internet Exchange Points (IXes or IXPs) to disintermediate ISP networks from sources of Internet bandwidth is a typical mechanism by which ISPs reduce their APBDC. Properly-sourced, bandwidth procured directly from its point of production in an IXP has a lower APBDC than bandwidth procured indirectly through an intermediary transit provider.[6][7][8][9][10]

Average Per-Bit Delivery Cost was first described under that name by Bill Woodcock in 2004, but built upon his previous work on Internet cost calculation, some of which was in collaboration with Andrew Odlyzko in the 1990s and with Zhi-Li Zhang and others in the early 2000s.[11][12][13][14][15]

References edit

  1. ^ Woodcock, Bill (July 19, 2004). "Average Per-Bit Delivery Costs". Packet Clearing House. Archived from the original on 2006-05-17. Retrieved February 4, 2012.
  2. ^ Woodcock, Bill (2018-04-20). "How does CloudFlare make money with their 1.1.1.1 DNS service?". Quora. Retrieved 24 June 2021. In ISP-economic terms, the greater the portion of their bandwidth has an Average Per-Bit Delivery Cost (APBDC) lower than their current overall average, the more they've succeeded in driving down their APBDC over time. To succeed, they need to drive APBDC down faster than their competitors.
  3. ^ IXPs reduce the portion of an ISP's traffic which must be delivered via their upstream transit providers, thereby reducing the Average Per-Bit Delivery Cost of their service. Archived August 29, 2006, at the Wayback Machine
  4. ^ "Software Defined Everything: SDx Infrastructure". SDx Central. Retrieved 24 June 2021. What SDx Applications Need from the Infrastructure: Efficiency: get more bandwidth, compute, and storage with fewer inputs (monetary or human); reduce average per-bit delivery cost.
  5. ^ CTU Bulletin on the Proliferation of Internet Exchange Points in the Caribbean
  6. ^ "Internet Exchange Services". Telehouse. Archived from the original on 2021-06-24. Retrieved 24 June 2021. An IXP reduces your average per-bit delivery cost by reducing a portion of your upstream traffic.
  7. ^ "Enabling Environments to Establish Successful IXPs". Internet Governance Forum. 2015-11-30. Retrieved 24 June 2021. When two local networks are exchanging traffic, an IXP can reduce the portion of a network's traffic that must be delivered via upstream transit providers, thereby reducing the average per-bit delivery cost of service and latency as well as improving routing efficiency and fault-tolerance.
  8. ^ "Mwanza IXP – MIXP". Tanzania Internet Service Providers Association. Retrieved 24 June 2021. The Mwanza IXP is operated by TISPA in order to provide a neutral peering point for IP networks in Mwanza, Tanzania. This allows for local routing of Internet traffic in Mwanza, between network operators, via peering connections, reducing the average per-bit delivery cost for Internet traffic, reduce latencies, increase available bandwidth and improve quality of service.
  9. ^ "Two Internet Exchange Points in Khuzestan, Isfahan". Financial Tribune. 2017-02-05. Retrieved 24 June 2021. The new exchange points will significantly reduce the average per-bit delivery cost of Internet services. This means that the price of the Internet for end-users will decrease significantly in the next few years.
  10. ^ Corbato, Steve (2020-04-27). "On Internet Exchange Points and Local Peering in the Time of COVID-19". Link Oregon. Retrieved 24 June 2021. IXPs enable local peering, which creates, in essence, a local network of networks and reduces the amount of an ISP's traffic that must be delivered via its upstream transit providers. And because traffic between local peers generally travels free, this can also help to reduce the average per-bit delivery cost while increasing efficiency and network resiliency.
  11. ^ Woodcock, Bill. "White Paper on Transactions and Valuation Associated with Inter-Carrier Routing of Internet Protocol Traffic" (PDF). Packet Clearing House. Retrieved 24 June 2021.
  12. ^ Woodcock, Bill. "Internet Topology and Economics: How Supply and Demand Influence the Changing Shape of the Global Network". Packet Clearing House. Retrieved 24 June 2021.
  13. ^ Odlyzko, Andrew (1998). "The size and growth rate of the Internet". First Monday. 3.
  14. ^ Woodcock, Bill. "Introduction to Exchange Point Economics". Packet Clearing House. Retrieved 24 June 2021.
  15. ^ Woodcock, Bill; Zhang, Zhi-Li. "A Straw-Man Pricing Model Addressing the Multicast Deployment Problem" (PDF). University of Minnesota & Packet Clearing House. Archived from the original (PDF) on 2021-05-02.