Add a figure for AWS architecture

[eparejatobes]

Using https://aws.amazon.com/architecture/icons/.

[laughedelic]

Check out https://github.com/ohnosequences/loquat/issues/63

[laughedelic]

Here's a draft:

Comments:

• the whole pipeline consists of several steps each of which is a loquat (see diagram in https://github.com/ohnosequences/loquat/issues/63, I'll add it here later)
• each step is independently scalable: you can configure type and number of workers you want to use
• split, merge and counting loquat "sizes" should be proportional to the number of input samples
• BLAST and assignment loquat "sizes" should be proportional to the sizes of the samples (i.e. number of chunks after split)
• this allows to analyse data as fast as you are ready to pay want: the most resource-demanding part is heavily parallelized

@ohnosequences/docs What do you think?

Other things that could be added to this diagram:

• example instance types (some steps require average hardware, others — more powerful)
• user interaction (launching each step, uploading input data, getting results). I didn't add it because it clutters the diagram quite a lot and it's more about data flow
• Flash step for illumina reads. It just doesn't fit 😅 (this diagram service has a limitation on the size of the grid for a free account). Also it's not clear how to show that it's optional. I would just explain it in text, because it's not a part of the core MG7 pipeline, more like a data preparation step.