| Teraoctets per second (To/s) | Yottaoctets per second (Yo/s) |
|---|---|
| 1 Teraoctet per second | 1 × 10-12 Yo/s |
| 2 Teraoctets per second | 2 × 10-12 Yo/s |
| 3 Teraoctets per second | 3 × 10-12 Yo/s |
| 4 Teraoctets per second | 4 × 10-12 Yo/s |
| 5 Teraoctets per second | 5 × 10-12 Yo/s |
| 10 Teraoctets per second | 1 × 10-11 Yo/s |
| 20 Teraoctets per second | 2 × 10-11 Yo/s |
| 25 Teraoctets per second | 2.5 × 10-11 Yo/s |
| 50 Teraoctets per second | 5 × 10-11 Yo/s |
| 100 Teraoctets per second | 1 × 10-10 Yo/s |
| Reference | Teraoctets per second (To/s) | Yottaoctets per second (Yo/s) |
|---|---|---|
| A dial-up modem | 0.000000007 To/s | 7 × 10-21 Yo/s |
| Typical home broadband | 0.0000125 To/s | 1.25 × 10-17 Yo/s |
| Gigabit Ethernet | 0.000125 To/s | 1.25 × 10-16 Yo/s |
| Streaming a 4K film | 0.000003125 To/s | 3.125 × 10-18 Yo/s |
The teraoctet per second is a unit of data transfer rate equal to a thousand gigaoctets per second, or eight terabits per second. Its symbol is To/s. It describes memory bandwidth inside the fastest processors and the aggregate throughput of large computing systems, rather than any link between separate machines.
Graphics and accelerator chips are the clearest example. A modern accelerator uses stacked memory placed alongside the processor on the same package, and the bandwidth between them reaches several teraoctets per second. That figure is what allows thousands of arithmetic units to be kept busy at once, and it is now the property that most often determines how fast a machine learning workload runs.
The reason bandwidth rather than arithmetic has become the limiting factor is worth stating. Processors have grown far faster at computing than memory has at supplying data, so a modern chip can perform tens of operations for every octet it reads. Any calculation that touches memory more often than that is limited by the memory, and most real calculations are.
Supercomputer interconnects reach this range in aggregate. The network joining thousands of nodes carries teraoctets per second across the whole machine, though no single link does. The design problem is to arrange the topology so that any node can reach any other quickly, which is why these networks are built as multi-dimensional meshes and trees rather than as simple stars.
To make the number concrete, one teraoctet per second would fill a large consumer hard drive in twenty seconds and transfer the entire text of every book in a national library within a minute. Nothing in ordinary use approaches it, and no external cable of any kind carries it.
The unit also appears in descriptions of storage arrays. A large parallel filesystem, spread across thousands of drives, can deliver a few teraoctets per second in aggregate to a supercomputer, and that figure is quoted as a headline specification because it determines how quickly a simulation can be written out and read back.
One teraoctet per second equals 1,000 gigaoctets per second, 8 terabits per second, or about 0.9095 tebioctets per second.
The yottaoctet per second is a unit of data transfer rate equal to a thousand zettaoctets per second, or eight yottabits per second. Its symbol is Yo/s. It is the largest transfer rate the metric system named for thirty years, and it stands at the point where the question stops being one of engineering and becomes one of physics.
The physical limits are real and can be stated. Any communication channel has a capacity set by its bandwidth and its signal-to-noise ratio, a result Claude Shannon proved in 1948. Pushing a rate higher means using more bandwidth, more power, or more parallel channels, and each of those has a cost that grows without limit as the rate does.
Energy sets the sharpest bound. Thermodynamics requires a minimum energy to distinguish one state from another at a given temperature, and although practical systems are many orders of magnitude above that floor, the floor is not zero. At a yottaoctet per second even the theoretical minimum becomes a substantial power, and every real system multiplies it by a large factor.
There is also a limit from the medium itself. A single optical fibre has a capacity ceiling set by non-linear effects in the glass, which grow with the light power carried, so raising the power eventually degrades the signal rather than improving it. Reaching a yottaoctet per second would require something like a hundred billion fibres running at today's records simultaneously, which is a construction problem rather than a communication one.
None of this makes the unit meaningless. It is properly defined, it converts by the same rule as every other, and it appears in discussions of theoretical limits and in complete tables of the prefix system. A measurement system that stopped naming quantities at the point where engineering stops would be less useful, not more.
Since 2022 the metric system has had ronna and quetta above yotta, so this is no longer the top of the ladder. That extension was driven by data quantities rather than by rates, and nothing in transmission has yet given a reason to write a rate above this one.
One yottaoctet per second equals 1,000 zettaoctets per second, 8 yottabits per second, or about 0.8272 yobioctets per second.