Sun Ray clients requirements
A Sun Ray implementation requires a reliable network with low latency and LAN quality of service.
It should be sized for a sustained throughput of at least 1 Mb/Sun Ray client. If you wish to be more conservative, then use 3-4 Mb/Sun Ray client.
- A single hme (100 Mb) maximum support can range from 25 up to 60 Sun Ray systems.
- A single ge (1000 Mb) maximum support can range from 250 up to 500 Sun Ray systems
Sun Ray sizing (Sun whitepaper)
The recommendations below are based on realistic configurations for two distinct types of workloads. These recommendations are intentionally conservative and should be considered a reasonable starting point. All metrics indicated are based on the use of Sun Ray Software 4 09/07.
Microsoft Windows User - Workload Type 1
A user leveraging the Sun Ray Connector for Windows OS, accessing a fullscreen Windows desktop session. This user's set of applications has minimal impact on the Sun Ray technology sizing calculation because the Sun Ray server does not directly execute the applications. Instead, the Sun Ray server is running the Sun Ray Connector for Windows OS and simply provides a high quality display connection to a back-end Windows desktop instance, where the applications are executed.
UNIX Desktop User - Workload Type 2
A user accessing a fullscreen GNOME or KDE environment from a Linux or Solaris OS. In this case, all applications that are run have a direct impact on the sizing calculation, as they are executed directly on the Sun Ray server.
| User type | User Quantity | Processor | Processor Quantity | Processor Cores | Processor Frequency | System Memory |
| Workload Type 1 | 75-150 | UltraSPARC T1/T2/T2+ | 1 | 8 | 1 GHz | 8 - 16 GB |
| Workload Type 2 | 75-150 | x86/x64 | 4-8 | 2 | 2.6 – 2.8 GHz | 16 - 32 GB |
Detailed Sun Ray sizing in SPARC environment (Sun whitepaper)
| Lighter usage:
| Average usage:
| Heavier usage:
|
| # of Average users | Sun Server | CPU UltraSPARC III 900 MHz | Memory | Light usage | Heavy usage |
| 25 | Fire 280R | 2 x | 2 | 35% more users | 25% fewew users |
| 50 | Fire 280R | 2 x | 4 | ||
| 100 | Fire V480 | 4 x | 16 | ||
| 200 | Fire V880 | 8 x | 32 |
Sun Ray Scaling & High Availability (Sun whitepaper)
Sun Ray Software provides load-balancing services across multiple Sun Ray servers in a high-availability (HA) group. It is important to consider the potential number of users each server may need to support should one server in the HA group become unavailable. Should this occur, best practices recommend the remaining servers in the HA group must be powerful enough to support that server's workload in addition to their own. For example, here are two possible ways to deploy a 200-user implementation:
Two servers sized for 100 users each (option A)
Here, it is recommended that each server should be powerful enough to support all 200 users, should one server become unavailable.
Four servers sized for 50 users each (option B)
Here, the load normally carried by the unavailable server can be shared among the remaining 3 servers in the HA group. In this example, the additional load is approximately 16 additional users for each remaining server (50 users divided by the 3 remaining servers).








RSS exports