Scenario taxonomy
This page has a more up-to-date location: {{{link}}}
Scenario Taxonomy
Introduction
This taxonomy addresses different scenarios that may occur in implementation of XO laptops and school servers. The taxonomy in its current form is simple and by no means comprehensive. Hopefully, this will change over time.
Use cases
These scenarios can be reinforced with use cases that provide a hypothetical example of use and often help in understanding needs and assessing constraints.
Some info on real deployment examples (details not confirmed):
Low and intermittent power in schools. 8 Mb/s to Data center. 150 students. From DC to Internet 256 Kb/s. Max 512 MB RAM available for school server.
Decision was to use wireless AP in school. School server for mesh, DHCP, and DNS only. Possible moodle or hosting later. Squid on server in DC.
- Peru - URL?
- Uruguay -
- Cambodia -
- Colombia - (coming soon)
Server components
- Backup
- Upgrade and configuration management
- Network Gateway
- Caching Proxy
- Authentication
- Content management
- Repository
Uruguay Server Use Case
- Platform : IBM x3105
- Processor : 1.6 - 1.8 GHz AMD
- RAM : 2 GB RAM
- HD : 160 GB disk (two drives, in RAID 1)
- Networking : 3 NICs (one WAN, and two LAN)
- USB : Six USB slots
- DVD-RW for backup
- Notes:
- The intent was to have a WiFi network inside the school and another for outside the school, but they are currently only using a single WiFi network. -These machines stop working at temperatures higher than 40C. (Crash, not nice power-off)
-The RAM size was chosen to better support Squid and Dansguardian.
-The small disk size is due to the costs of disks when bought through server vendors and there not being XO backup at this point.
-Watchdog and remote monitoring boards are recommended (but not currently installed.)
Infrastructure constraints
Infrastructural constraints are very important and often drive the decision to use a laptop/mobile or desktop/lab environment or some combination thereof. Two major components of the infrastructure are power (grid-based electricity) and backhaul (connectivity to the Internet). In the following table, we have a combination of the two on a granular scale akin to low, medium, and high.
List of relevant variables :
- Server HW (RAM/Disk) - big, medium, little, none
- Bandwidth to Data Center - big, medium, little, none
- Bandwidth to Internet - big, medium, little
- Power in school - enough for AP and 2 x servers, enough for one server max, not enough for any servers
- Wireless access - Wireless AP, active attenna, none
More detailed break down of these details at: Networking scenarios
Size of school - 10 - 50, 51 - 100, 100 - 200
The idea is to pick your available resources on all the metrics that you know listed above. Then we can quickly say what services you can use and identify any constraints within those variables.
e.g. If school greater than 150 student, then must use wireless AP instead of active attena. If zero BW to data center, then peer to peer mesh only (or maybe still use server and attenna for local mesh?) and sneaker net to internet.
Or in reverse starting from requirements.
e.g. If 100 students and need local mesh and browsing of big libraries in data center and some browsing of Internet then must have lots of power, server with 1GB RAM for Squid and 2 x active attennas and >1Mb/s to data center and >512Kb/s to internet.
See also deployment guide underway at:
etc..
Gregorio 09:15, 27 March 2008 (EDT)
| No grid infrastructure | Grid infrastructure available, but does not work reliably | Reliable grid infrastructure | |
| No backhaul | Worst case scenario | ||
| Limited backhaul (cellular, satellite, dial-up) | |||
| Good backhaul (DSL, Cable, or anything that can be called “high speed”) | Best case scenario |
Note here we also have to consider the different approaches that a NOC can take in each country. This is hard to tabulate (at least now for me, says martin ;-) ) but we'll eventually have a grid describing it.