Public Sector is facing the biggest changes in the history of information technology. From computing to networking and secure storage, technology traditionally has required fixed, inflexible hardware environments.  Server virtualisation was the first step towards changing that model, offering new flexibility for using computing resources more efficiently.  Virtualisation in the data centre for workload mobility brought efficient resource utilisation and high availability and helped to drive down cost and those benefits are now being realised by Public Sector in reducing spend with and of course increased competition by service providers to maintain revenues.

Many of those large infrastructure contracts have been extended to the maximum term and there is a genuine desire to create market opportunity with disaggregation of the supply chain as a key strategy. The next generation of major infrastructure programs are now underway with programs in Devolved Governments Super Councils and even with Crown Departments such as HMRC, DWP, MoD and even Health (the requirements for Social Care Network (HSCN)) to name but a few.

We are seeing a major shift towards cloud computing for public and private cloud applications, it’s clear that the networks deployed must also keep pace with requirements for faster, more flexible connectivity availability to the smallest location and finally the ever-increasing demand to meet the new round of budgetary constraints which for many suppliers will become apparent in the next two years.

 

Traditionally wide area networks (WANs) and to an extent the Public Services Network (PSN) connectivity was deployed was built around data centres. The “Authority” typically purchased an MPLS service (for example a PSN connectivity solution) from a quality proven carrier.  The carrier then managed all the complexity of building a large-scale WAN.

As the user starts looking more towards the cloud both public and private for a range of applications, the most important change required from the network is flexibility since the cloud is about on-demand service, elasticity, and ubiquitous access.  Today, in the PSN market such flexibility using an MPLS based solution can be costly to deploy or to manage. Either network vendors start cutting significant costs (over and above the circa 20% saving in the last two years) and improve automation to enable on the fly changes to the network using simple tools; Or the PSN now needs an alternative approach to drive commercial operational and innovation behavior.

Those that helped create the PSN standards who invested in this market should feel justifiably proud that the PSN has met many of the original objectives set by the Cabinet Office. Regardless of the elevator pitch of a network of networks many of the solutions deployed and used are delivered by only a handful of suppliers (less if you consider wholesale arrangements) and the vibrant market place envisaged by now of twenty plus suppliers has not been achieved… yet.

 

This author does not advocate a rip and replace of the PSN, but rather the migration of the PSN to support a truly disaggregation supplier and connectivity model.

 

Examples already exist such as the Obligations Framework published for HSCN which replaces many of the costly aspects for suppliers when delivering network connectivity. Better HSCN was proposed and agreed by a wider supplier community than the original PSN supplier community. Checks and balances and commercial impact on the end user over the standards proposed was ensured. A role managed by Innopsis in consultation with the NHS Digital team.

 

With changes in policy being debated by GDS and thought leadership becoming apparent GDS are suggesting networks need to meet the demand for elastic service with ubiquitous access. To support this, they have stated the capability of the Internet for the Authority should become part of their WAN strategy or be considered. I suggest this could be tweaked and that the user should be agnostic to the connectivity they consume provided the functionality is delivered and suppliers and users should be free to innovate.  But with the Internet, and use of any wider transport capability security, availability, quality of service and common gateways between suppliers will rightly is one of the biggest obstacles.

 

At a conference hosted by Innopsis regarding the future of PSN I outlined a vision for the next generation PSN which converts the traditional MPLS based PSN network into a network of standards that operate based on an industry agreed Obligations Framework. The service should provide all types of cost effective connectivity, allowing the vendor to choose how the network functionality is provided. But mandated each supplier will create an interconnect at nominated carrier natural datacenters (similar to HSCN but enhanced as this will include peering with the legacy PSN). This new network would provide the end user a wider choice of connectivity solutions, the ability to meet changing demand and supports onward connectivity to cloud services from the peering interconnects providing onward savings for the user.

 

The ability to provide this functionality already exists with technology such as Software Defined Networks (in this case SD-WAN) which also provides cost effective solutions to the issue of security over any connectivity. Today the PSN struggles to offer cost effectively and if a user has more than 1,000 locations it can be technically challenging.

 

For this new service, I suggested we should mandate the following: measurable reliability, real-time performance and adherence to government grade security policy.  A thought for you the reader, with the advent of the cloud, flexibility has become equally critical, traditionally, since networks were built around physical topologies with policy and configuration set at every hop, every network element had to be configured when a policy change was required, from routers to switches to firewalls.  This made the network very rigid and required multiple provisioning steps to make any change.

Cost effective IT is a trend that cannot be ignored or constrained by the network it requires the network (unless we are to witness a paradigm shift in current market conditions) towards overlays, where the physical topology of the network does not determine the traffic flow from the user to the application. Instead, virtual overlays you build based on policy define the traffic flow.

 

The PSN of the future created around an Obligations Framework and support for vendors to agree to meet those requirements will allow the end user (Authority or its service provider) to build arbitrary virtual topologies based on application requirements, each line of business can have its own topology and even choice of network connectivity used.  For example, financial applications should only connect to the data centre, so they can use a hub design, while voice should be fully meshed.  Applications should have visibility to all circuits (MPLS, Internet, and broadband), and based on SLA, the best path should be selected.

 

Today Public Sector in the UK is consuming cloud applications at large scale. Cloud is the largest VPN we will build based on arbitrary connection points. We are solving the problem of user mobility with laptops, thin client devices smart phone and tablets. Now it’s time to solve the problem of work load mobility. Just like users, work load now can be provisioned at any location. The PSN evolution requires the network to be cost effective, competitive and open to all supplier secure, elastic, and extensible.

 

 

Michael Bowyer

Loddiswell Consulting

© 2017 Loddiswell Consulting Limited

%d bloggers like this: