Infrastructure review synthesis: Post 5, Servers and storage

This post is the fifth in a series of ten posts that have been created to identify the best practice found in FE colleges by the Jisc infrastructure review service. An introduction to the infrastructure review synthesis project is provided in the first post in the series.

Servers and storage

The vast majority of colleges have virtualised their on-premises server infrastructures. The effective use of virtualisation means that the total hardware footprint has been reduced in most cases, with a small number of server hosts running a virtualisation hypervisor that in turn runs a number of virtual servers.

Virtualisation is a useful technology as it abstracts the services from the hardware on which they run, it also makes backup and restoration faster and more reliable. The most mature colleges have sufficient compute resources, notably the best colleges ensure that an element of the can system fail, and there continues to be sufficient compute to manage without disruption. This means that all systems and services can stay online, should any one host fail. The best colleges ensure that server virtualisation systems continue to be kept up to date, this ensures that future virtualisation and operating software updates will be supported by the hardware. In the best cases, compute (server host) hardware is fully under manufacturer support and can support current virtualisation software or server operating systems.

The most secure IT operations only use fully supported server operating systems and have eliminated any legacy systems or applications that only run on older operating systems.

The most advanced examples include development or testing environments to allow IT teams to test new systems or services or to simulate upgrades and other large-scale changes of systems prior to applying such to the production environment.

In the best examples server room contingency is outstanding with redundant UPS (uninterruptible power supplies), air conditioning and fire detection and suppression systems in place. In addition, the most resilient organisations include environmental monitoring and alerting inclusive of temperature and moisture sensors. Again, the most advanced server rooms are physically secure with good use made of electronic access control systems to both control and log access in place. In addition, CCTV systems or master key systems may also be used.

Most colleges have sufficient storage capacity, the best plan ahead to ensure that future growth can be accommodated alongside considering how to support media courses and other curriculum areas that make greater use of storage over other courses. In the best cases storage is being proactively managed to ensure that archiving and retention policies are enforced. In the most outstanding examples data protection officers are actively engaging with IT teams to ensure that policies are being technically applied.

The most resilient organisations make use of an active – active cluster with at least two main server rooms that are sufficiently specified to run the entire college enterprise from either location should one be compromised.

Alternatively, some colleges have chosen to implement a DR (disaster recovery) environment that would enable a sub-selection of vital systems to be run from a smaller infrastructure whilst the primary systems were brought back online. This is noted to be a less costly option over an active – active arrangement.

We note that take up of public cloud hosted IaaS (infrastructure as a service) is very low. We understand that this is down to IT managers determining that it is generally not cost effective for most FE colleges to make effective use of IaaS due to scale, high cost, and limited connectivity arrangements such as not having a resilient internet connection.

Also, due to the specific use case of college based systems such as the fast storage or low latency required by media-based courses such as games design, photography is not in all cases, appropriate to make use of offsite storage as this could result in poor performance. It is important to note that whilst IaaS use is very low, SaaS applications are in increased use as is explored in section 6.

How Jisc can help:

  • The infrastructure review service has enabled IT teams to understand how to upgrade or improve existing compute and storage systems in order to make the best of what equipment is already available or to make useful decisions about where to target future investment.
  • The applications review enables the future infrastructure requirements to be well understood by the team so that applications may be provided from a SaaS, or on-premises method.
  • Most colleges have already moved core applications to SaaS delivery, however there are a number of opportunities to streamline service delivery, such as trust and identify management and single sign on, and these have been highlighted through the infrastructure review service.
  • The Jisc cloud consultancy service offer enables colleges and other organisations to obtain support to move applications and other systems from on-premises to cloud service provision.

Leave a Reply

Your email address will not be published. Required fields are marked *