System Resilience and SustainabilityThe HELIO infrastructure has been implemented with a service-oriented architecture; in this the tasks that need to be performed are implemented as independent services. We have used features of the architecture to create an infrastructure that is extremely resilient. By the same token, this is intended to also ensure the sustainability of the infrastructure.
ServicesThere are multiple instances of most of the services – these are deployed in institutionally and geographically separate locations. Configuring the system in this way means that in the event of a service being unavailable an alternate instance can usually be found; this increases the reliability of the system and ensures that there is time available to resolve any problems.
A snapshot of the instances (and locations) of the services that have been deployed is shown in the table below; the most up-to-date list of the services that are available can be found in an Overview page in the HELIO Registry (HRS).
|Service||Location of Instances|
|HEC||UCL-MSSL (2), VO Paris|
|HFC||VO Paris, UPS/IAS|
|ICS and ILS||UCL-MSSL, UCL-Phys, VO Paris, UKSSDC|
|DPAS||UCL-MSSL, UCL-Phys, VO Paris|
|HPS and HSS||TCD|
|Access||Location of Instances|
|HRS||UCL-MSSL, UCL-Phys, VO Paris, UKSSDC|
There are also several instances of the Registry that are used to manage access to the services. The IP address to the prime Registry can be changed by amending the DNS record although the HFE and IDL/SSW code both know the locations of the registries and derive all address information from the Registry that they are able to contact.
Data SetsThe resilience extends to the data access capabilities of HELIO that provide the user with a degree of redundancy.
- We have included multiple sources for some popular data sets; if the site that is normally used is not available then the next one is tried.
- If there are multiple sources, the user can express a preference as to which to use.
- For data sets that are sourced from NASA's Virtual Solar Observatory (VSO), there is an automatic failover between instances of VSO should the one normally used become unavailable.
Last updated: 24th April 2015