Warning: Cellular Bolt-on Models May be Harmful to Nets

November 12, 2013

OPINION by Art King, SpiderCloud Wireless, Director of Enterprise Services & Technologies

When the partner of a mobile operator or business representative shows up with a cellular bolt-on module to attach to the existing Wi-Fi access points, to make life better for all mobile devices operating indoors, it may sound like a win-win proposition.

But, for enterprise IT professionals who have been exposed to the challenges of indoor cellular, easy is like waving a red flag in front of a bull. If the Wi-Fi network is a scalable WLAN with controllers and access points, how can 3G simply be strapped on without consideration for security, scaling and interference issues? As the saying goes, “they don’t know, what they don’t know.” There are hidden challenges.

Before we discuss the challenges, keep in mind this snapshot of IT inside the enterprise:

Enterprise IT networking is overwhelmingly complex with pressure to “do more with less” and increased workloads.

Capital budgets are not keeping up with infrastructure refresh demands.

IT often ends up doing bare-bones technology implementation of new projects.

IT is burdened with management and operations of an array of software and hardware, all of different vintages, and supplied by various vendors.

IT has to deal with staff acquisition, training and retention in intense competition with other enterprises.

There is constant pressure on security and the integrity of the enterprise network.

Enterprise IT Bolt-on Concerns, Challenges

Operator Trust Relationship:

It starts with trust. Because the bolt-on cellular coverage and capacity network must be integrated into the live Wi-Fi infrastructure, the enterprise must trust that the business partner/mobile operator has the expertise and capability to build and support the implementation. Can an enterprise IT leader stake his or her career on the business partner/mobile operator with a bolt-on deployed inside the perimeter? If a trusted relationship with enterprise IT has not been established, the bolt-on should not be considered by IT.

Capital Planning and Budgeting:

The bolt-on cellular module requires an access point that can host the bolt-on and PoE+ capable Ethernet switches. If the required hosting AP and PoE+ are not widely deployed, this presents a hidden capital problem. For the enterprises that want to implement the bolt-on on a large scale, they will need to replace older APs, possibly the WLAN controller and Ethernet switching. If the bolt-on cellular network is competing in a tight capital environment, it could take 18-24 months to get the funding and implement the basic foundation to connect and enable the bolt-on to be installed into the access points. To further complicate the approach, writing off equipment in the middle of its life has dire consequences for future budget requests. Enterprises are accustomed to a five-to-eight year depreciation and replacement cycle. Shortening the life of already invested equipment has budget and career implications.

Legal Indemnification:

Because each bolt-on” cellular module is attached to production network equipment, will the mobile operator indemnify the enterprise for breach of network security, should a device sniff live network traffic? And, will the mobile operator ask the enterprise to indemnify them from any damage caused by a breach using the IT network that may impact the enterprise network?

Context: SpiderCloud Wireless has deployed separately from the enterprise’s production data network in a VLAN and used third party audit of the architecture conducted to provide external assurance of integrity to the enterprise.

Operations Management, Ownership, and SLA:

In a mixed ownership model (where the WLAN is managed by enterprise IT and the cellular bolt-on is managed by the operator), day-to-day operations and SLAs need to be clearly defined. A mobile operator’s SLA can be impacted by actions of the network engineers in the enterprise. For accounts where IT procurement determines SLA agreements without enterprise network operations, the operation’s responsibility gap for the bolt-on cellular network inside the enterprise’s premise could impact long-term operations. This is about trust, and clearly defined lines of operations and responsibilities.

Information Security Acceptance:

If a cellular bolt-on module or network is deployed behind the firewall of an enterprise, there are a plethora of systems integration and operations hurdles to consider. Even if the WLAN network is owned and operated by the mobile operator, the customer will have to accept that all of the mobile operators’ subscribers may attach to the system on the same network with enterprise data. This is simply unacceptable to information security policies of military, government, financial and other security conscious entities. Where enterprise IT demands a physical separation, a bolt-on will most likely be eliminated from consideration.

Network Engineering:

A bolt-on cellular network approach is deeply intermingled with the customer’s existing deployed architecture. To take it into service will require edge-to-edge reconfiguration/replacement of Wi-Fi access points, the Ethernet switches, and possibly other components. This implies a significant amount of project management, IP network engineering staff resources, planning and after hours implementation work to prepare:

Access points each need to be taken down to install the bolt-on (considering that the existing access point can indeed power a strap on module in the first place)

Most likely, all older access points will require a new AP replacement. The financial ramifications are substantial to the enterprise and/or mobile operator.

Access Point Location Mis-match:

Protocols such as 802.11A, G, N, AC, AD and future types each have their own propagation characteristics for RF. Power outputs and frequencies are different between Wi-Fi types and the variety of bands available in the mobile operator’s licensed spectrum. The site survey process to add a bolt-on cellular module or network to existing APs/WLAN is more complex than doing a Greenfield installation of integrated Wi-Fi/Cellular small cells because the existing AP locations become a major factor for installation. The bolt-on could require substantial additional expense in cabling due to the shuffling of access points to meet cellular needs – causing disturbance of coverage patterns that the enterprise’s Wi-Fi users have become accustomed to. In summary, re-locating access points can quickly make the business case go negative for all parties involved.

Software Release Compatibility:

For a given mix of enterprise hardware and feature requirements, there may be only one release of operating system software that supports a particular enterprise’s mix of Wi-Fi operating requirements. This is a “hidden in plain sight” fact that senior network engineers are well acquainted with, which can influence infrastructure evolution. Another factor to consider is that most enterprise LAN configurations are unique to the enterprise. With this in mind, operating system revisions needed to support the bolt-on must be regression tested and qualified against the present usage. If the tests show a negative impact on the enterprise Wi-Fi operating environment, a cellular bolt-on will not be accepted for use. Where software intersects, plan for the unexpected conflicts.

Hardware/Firmware Compatibility:

This is similar to software compatibility, but may require replacement or re-flashing of the access points. Replacement is a physical activity that would become the path to installing a bolt-on onto the access points. Regression testing and multi-access dependencies as outlined in software compatibility, apply here also. Where hardware intersects, plan for the unexpected conflicts.

Options for Multi-Access RAN Synergies

Will a bolt-on cellular network enable future capabilities and services of the merged 3G + Wi-Fi network? Can the bolt-on cellular modules use the existing AP as a source of power and transport access? If the direction of a bolt-on cellular network is ‘a’ with no possibility for future services, perhaps the operational complexity of adding it to the existing APs might have little incremental value, but carry a lot of overhead in the life of the installation. In summary: If the bolt-on only uses the access point for power and transport, stop and think about future ramifications.

So what about Return on Investment (ROI)?

At first, the bolt-on proposition is compelling. It is positioned as an easy way to leverage the massive investments that enterprises have made in Wi-Fi. The low amount of access points in production networks today that can actually support a “bolt-on” module are few. Therefore, is the real driver for this technology not so much a “bolt-on” cellular module for cellular coverage, but rather the replacement of older Wi-Fi equipment installed in enterprises?

Peeling back the layers of the onion reveals complex enterprise issues that most mobile operators are not geared up to address. Holistically, Wi-Fi is a potential enterprise revenue generating opportunity. A bolt-on cellular network to Wi-Fi network is only a limited indoor coverage and capacity solution for small businesses. If the mobile operator is seeking to resolve indoor coverage and capacity for larger enterprises, the systems and operations integration and RF hurdles may quickly kill the business case. The integration hurdles for a bolt-on implementation for medium to larger enterprises requires careful lifecycle cost analysis.

Will a bolt-on go-to-market strategy fail because every customer that wants it will have to enter a capital budget cycle to get refresh funding for its existing WLAN? What about scaling and interference issues for the cellular “bolt-on” modules? How will that be handled? SpiderCloud Wireless has already covered these issues in this network architecture paper.

A Safer, More Scalable Option for Medium to Larger Enterprises

A complimentary LAN-RAN strategy has few of the complex engineering, technology, and operational dependencies that we believe will be a barrier to success of overlaying a cellular-to-Wi-Fi bolt-on architecture on top an enterprise’s incumbent Wi-Fi infrastructure.

SpiderCloud E-RAN enabled mobile operators can offer their customers a flexible cellular solution that can be loosely coupled to customers LAN via an isolated private VLAN (see “Using VLAN’s in Network Design” by Nexus). The other option is a physically isolated network from the enterprise LAN infrastructure using a separate Ethernet LAN. With deployment costs lower and implementation agility far higher, a SpiderCloud Wireless E-RAN architecture is IT friendly, solves the coverage and capacity problems and has a seamless path to future opex services.

A bolt-on module may save a cable installation charge per access point for a small business deployment, but introduces a new family of unknown expenses and issues. A strap on cellular module or network is not suitable for medium to large enterprise networks.

Therefore, a cellular bolt-on proposal should come with an enterprise IT health warning. “Use of a bolt-on may have unforeseen consequences to your network, and may be unsafe to your IT career.”

But, what do I know? I was only responsible for a global brand’s enterprise IT mobility and wireless issues for 10 years – now working to help educate mobile operators and enterprise IT, as we all go mobile.

Twitter: @EMobilityInside

Visit our Enterprise IT site http://SpiderCloud.com/EInsider