Putting The Network In A Box
4th December, 2018
We’ve Come A Long Way
Parking up my car to visit my local town centre, I walk right past the old telephone exchange building. I spent a lot of time in those buildings, awestruck at the vast amounts of equipment deployed to deliver fixed narrowband voice services to the masses. It is amazing to think that now, not only is a lot of that building empty, but the population masses, which were so dependent on them for those restricted narrowband services have access to vastly superior services, delivered over broadband and mobile networks.
In this respect, 4G (LTE) has proven to be one of the most significant innovations. Not only does it deliver broadband services over mobile networks, unleashing an unparalleled tidal wave of application layer innovation, to the benefit of hundreds of millions of users, but it has done so over an attractive flat network architecture, which implements very complex network functions across interfaces and using protocols, which have proven to be well defined and very consistent.
Whilst the devices we now use to communicate over 4G networks are small, the networks themselves are certainly not. Those same masses who are now so familiar with the benefits of mobile broadband are equally familiar with the vast quantities of infrastructure required to deliver those benefits. Masts on top of a base station are a part of the landscape and even if the general public was not aware, telecommunications veterans would have some knowledge of the room sized transmission systems, mobile cores and support systems required to support these base stations, all contributing to the view that whilst the 4G device was small, the 4G network was not.
The Unsolved Problem
In fact, the only thing actually mobile in a mobile network, is you the user and this creates issues, the most obvious of which those same masses are certainly aware.
In a mobile network, coverage is dependent on the nearest base station and the location of the base station is dependent on the mobile network operator.
What can be frustrating to the typical user can be devastating to others. For organisations such as the emergency services, or the military, reliable communications really are a matter of life and death. So much so in fact that historically, they have eschewed the benefits provided by the mass market mobile technologies in favour of non-3GPP solutions.
However, thanks to 4G, its ability to deliver broadband over mobile networks, open ecosystems supporting mobile applications and in particular the 4G flat network architecture, there is a solution now available to the emergency services and to the military, which allows them to share all of the benefits of the 3GPP standards and 4G in particular, without having to suffer the coverage issues which continue to frustrate mass market users, particular in rural locations.
The problem is solved by changing the rules of the mobile network.
We are starting from the premise that network is statically defined, with coverage dependent on your proximity to base station locations defined by people who were designing for a model based around their definition of average usage and average users.
What would happen if the network itself was mobile, if it could be conveyed to the location where coverage is required to support essential communications, with the guarantee that neither coverage nor capacity would be an issue, due to dedicating the network to a closed community of emergency service or military users ?
This is the capability delivered by Network In A Box (frequently abbreviated to NIB, or NIAB). Rather than using the statically defined macro network to provide coverage according to somebody else’s rules, create your own micro network and deploy it where you want, when you want it, according to your rules.
The concept is simple – package all of the components required to create your own portable bubble of coverage. However the realisation is pretty sophisticated.
Creating The Network In A Box
Critical to the operation of the Network In A Box are applications. Network In A Box products will implement the standard data transmission capabilities of traditional 4G networks. However, the emergency service and military communities have specific requirements over and above the mass market standard.
The 3GPP standards process has defined a number of “Mission Critical” services to cover voice, data and video services, for emergency service and military usage and it is possible to retrofit these into a public network. Similar services are required in a Network In A Box environment. However, it needs to be remembered that all of the users of the Network In A Box are high priority, operating from a well defined closed user community and it could be argued that the “Mission Critical” elements, based around priority and pre-emption concepts, are actually of less interest in a Network In A Box environment than in a public network.
Mission Critical Voice for example can be readily implemented across mobile broadband networks by third party Push To Talk applications, available today and not dependent on 3GPP standards driven Mission Critical upgrades. That is not to say that Mission Critical functionality should not be deployed on the Network In A Box, rather that the deployment of the Network In A Box itself can still be achieved without the Mission Critical functionality being present.
And what of the Network In A Box platform itself ?
The starting point is the base station. The requirement is for a compact device which will provide coverage to a specific community of users, over a distance which would typically be less than you would require from a macro network. Requirements in fact which are ideally suited to small cells.
Small cells are radio products with much smaller form factors than traditional macro base stations. They have been created specifically for the purpose of providing fill in coverage and capacity and for specialist networks. The size reductions achieved along with environmental and safety enhancements allow small cells to be deployed in a wide range of scenarios where macro base stations are completely impractical :
- Mounted on vehicles
- Deployed on drones
- Carried in backpacks
Small cells, deployed in these scenarios and with the addition of an appropriate power source, cables and antennae, deliver the basic component for addressing coverage and capacity. However, to guarantee operation in a remote environment, the small cell needs to have access to a core.
The great thing about 4G is that it has been created with small cells in mind. Both 2G and 3G are hierarchical networks, there is a BSC (2G), or RNC (3G) between the base station and the core network. Moreover, neither has been created with the intention of supporting small cells. 2G does not really accommodate small cells at all and 3G does so by introducing another component, the FGW, to replace the RNC in the topology, changing the interface between the “radio” and the “core”.
The Core Challenge
4G, with its flat architecture, does not discriminate between macro base stations and small cells. The interface between the radio network and the 4G core network (Enhanced Packet Core – EPC) is the highly robust and well defined S1 interface regardless. This is of real significance in the Network In A Box environment because it creates the possibility of off shelf 3GPP compliant 4G core products interfacing with the small cell in the same manner as they would with macro base stations. However, there are some real challenges.
Firstly, in order to implement a portable solution, you need a core network product which can run on a hardware platform small enough to be carried in the vehicle, drone, or backpack.
Secondly, you need a product which can be cost effectively deployed in large numbers, with each core supporting a small network.
Finally, you need a high quality, flexible, scalable and easy to use product which can operate in very challenging environments.
Legacy core networks have been created as one off instantiations deployed to support extremely large networks. They come with a reputation for massive feature support and the ability to operate networks supporting tens of thousands of base stations and many millions of users. They are big complex products, running on massive hardware platforms, typically deployed once in an operating centre and then upgraded many times over their operational lives. They are designed for this purpose, are expensive to deploy and expensive to maintain.
The ideal hardware platform for the core network deployed as part of a Network In A Box is the radio chip, or possibly a coprocessor on the radio itself. Chip vendors such as Cavium have made product with enough processor cores, enough CPU power and enough memory to enable third party applications to be deployed without impacting on the fundamental performance of the radio.
However, although this excess of processor capacity is ideal for hosting a Network In A Box core, legacy core products are completely ill suited for this purpose. Considerable functionality developed on the legacy core network to meet the requirements of the Mobile Network Operators is just not required in the Network In A Box environment. Typically, they are too demanding in terms of CPU power and memory to be of any use in this type of environment.
The Network In A Box core product will be deployed in many separate instantiations, perhaps one per police car, or one for each army platoon. It can not be priced in the same way as the legacy core network. It is an environment in which you are looking for a product oriented around a “lots of small and inexpensive” concept whereas the legacy core has been created to fulfil a “one off, big, expensive” function.
Typically, the legacy core network products used to run the mass market networks supported by the public Mobile Network Operators will be deployed in a data centre environment, running on highly robust, easily accessed, easily swapped and easily upgradeable servers. Whilst there will be expectations of high performance from the software, especially in areas such as billing, it is a forgiving environment for software upgrades required to address bugs.
In contrast, a Network In A Box may well be deployed in a very unforgiving environment. On a battlefield, or supporting operations around a public safety emergency, quality is essential. There is no room for product failure in this sort of environment, lives depend on reliable operation and the ability to make software upgrades is extremely limited. Users of the product are looking for ease of use and predictable performance.
Legacy core networks are not the solution for Network In A Box.
The Right Core
The Network In A Box needs an alternative to the legacy cores and AttoCore has developed AttoEPC specifically for this type of application.
AttoEPC is a specialist, forward looking 4G core, implemented to the same exacting 3GPP standards as legacy cores, but with a number of characteristics making it much more suited to Network In A Box than legacy cores. AttoEPC does not support any 2G, or 3G functionality and has no circuit switching functionality. The most obvious first impact of this of course is that the software executable itself is much smaller – it does not need to run on top of the line, high performance servers, but can easily run in the more constrained hardware platform available within the small cell.
Focusing on 4G also means that AttoEPC implements only the functionality required by the application and does this to a very high standard. A development philosophy underpinned by continuous execution of thousands of automated test cases means that the software is released to the highest standards of quality. With less to go wrong, issues are captured and can be addressed immediately and the test case suite is enhanced whenever new problems are detected.
Such a commitment to quality makes AttoEPC ideally suited to deployment in the hard to reach places where Network In A Box is most often found. However, quality without ease of use is of little relevance and so the ability of AttoEPC to tick both boxes is a major advantage.
The pricing of the AttoEPC is based around expected deployment in specialist applications at the edge of the network, where the key characteristic is cost effective deployment of multiple individual EPCs, rather than the support of very large public networks from within a data centre, where pricing assumes that a large scale day one product will grow larger over time.
AttoEPC has also followed an “open” approach to development since day one. Implemented on a Linux platform, AttoEPC has been created with support of third party applications a principal consideration. One such application of course is Push-To-Talk, essential to many Network In A Box deployments and AttoEPC is already proven to interoperate with a number of existing Push To Talk applications. Looking ahead, in addition to working with the existing generation of well proven Push-To-Talk applications, AttoEPC will also embrace the Mission Critical functionality coming out of the 3GPP standards programme.
AttoEPC is ultra-scalable, allowing cost effective support of the very smallest Network In A Box deployments and is sufficiently flexible to be deployed embedded directly on the radio chip, or on a small coprocessor, or standalone server.
Whilst implementing a discrete bubble of coverage as a basic capability, communications between the Network In A Box and both the internet and the macro network can be achieved using standard interfaces. Of more interest, however, is the ability to “mesh” together stand alone Network In A Box “bubbles”, to allow communications between users on “bubble A” and users on “bubble B”, without the need to transit a classical macro network. AttoCore has implemented this functionality using patented technology, which also allows users on “bubble A” to migrate to “bubble B”, without losing context within the Network In A Box environment. Such capabilities have significant application. A typical scenario may include the ability of platoon commanders cut off from company level intelligence to exchange information on the battlefield.
A major requirement for both military and emergency services applications is the ability to pin point the location of Network In A Box nodes and service users communicating with the Network In A Box. In a search and rescue scenario, this capability could allow a first responder to exactly define the location of a missing or injured person, simplifying the process of dispatching specialist support to the right location. In a military environment, it could be used by a platoon commander to locate threats, or to maintain a full picture of the location of his own troops. These are the sort of capabilities enabled by AttoCore’s AttoLCS Location Based Services, deployed with AttoEPC as an enhancement to the basic Network In A Box functionality.
What Next For Network In A Box
A Network In A Box is a coming together of a small cell, a fit for purpose core and specialist applications. AttoCore has implemented its AttoEPC to be ideally suited to deployment in Network In A Box and has worked with a number of small cell vendors, systems integrators and applications developers to create Network In A Box propositions for markets around the world. AttoEPC is already deployed in Network In A Box applications with the UK emergency services and will shortly be deployed in trials in the USA. Other partners are working on proposals for deployment in Europe and in Asia.
The result of all of these endeavours is something which is quite amazing. The principal mobile broadband capabilities implemented by 4G networks comprised of thousands of base stations and highly sophisticated mobile cores, but realised on a much smaller scale in a single portable box, which can be deployed in a car, on a drone, or carried in a backpack.
It is a concept which takes the 3GPP standards out of the public mobile networks and expands them into domains regularly considered inaccessible. Whilst principal usage is currently anticipated from the emergency services (including search and rescue) and the military, there is expected to be a lot of other application, particularly as the concept of unlicensed LTE takes hold.
Unshackling the base station, embedding a specialist core such as AttoEPC and evolving it into its own stand alone portable network is a reality which could bring significant benefit to many different user communities.
That reality could not be more removed from the static narrowband networks implemented inside our old telephone exchange building !