Sharing a Softswitch

I see a lot of companies buying softswitches and it makes me wonder if it might not be a better economic idea to share a switch with somebody else. There is so much potential for savings that anybody thinking of installing one should consider it. In the following I will discuss what it means to share a softswitch and look at the pros and cons.

The Basic Requirements. Sharing requires that somebody already owns a softswitch that can control geographically separated gateways. Such a switch must be able to share both the inter-machine trunking gateways, which facilitate PSTN interconnection and the media gateways which facilitate interconnection to customers. Sharing is also going to require interconnection between the two sharing parties. This means there must be some sort of trunking established between the two parties that can be either traditional TDM trunks or IP trunks.

What Would be Shared. The following elements are the components of a softswitch that can be shared, meaning that only one of each of these needs to be purchased:

  • Core softswitch – the device that contains subscriber and routing information;
  • Feature Server – the device that facilitates phone features;
  • Session Border Controller – the device that provides security between the softswitch and the gateways, some CPE and other networks;
  • Signaling Gateway – the device that interface with the SS7 network;
  • Unified Messaging or Unified Communications server – the device that stores, controls and converts data used in unified communications services like voice mail, email notifications, voice to text and text to voice, etc.

What the Second Company Needs to Buy. If you are going to share somebody another softswitch you only need to buy a few components. These are referred to as the distributed elements in a softswitch network. The two elements that you must buy are:

  • Inter-machine Trunking Gateway – for local interconnection to the PSTN. This lets the lessee still connect to the world using the same connections in place your legacy switch;
  • Media Gateway – for local connection to local distribution network and/or CPE.

Data connections required. As mentioned, there must be a data connection established between the softswitch and the new location that allows communications between the shared elements housed at the core softswitch and the distributed elements found at the new location. The size of the data pipe/connection required depends on the amount of data required between the shared and distributed elements. It only requires a few megabits per second to transmit voice traffic. You’ll need a connection in the tens of megabits per second if you are using a lot of features like voice mail & unified communications or a lot security invocation.

Partitioning. The owner of the shared elements will have to ‘partition’ the shared elements in order to ensure that subscribers of the sharing company can’t be seen or manipulated by other sharing companies; Partitioning will also hide the call detail records, routing control and will make sure that the shared elements can communicate through to the multiple networks operated by the sharing companies.

How to Determine if Sharing is the Right Idea. The big benefit is the savings, but there are other considerations:

  • It is inexpensive to buy only the distributed elements. Depending on how large you are, the distributed elements could cost anywhere between $25k and $200k which is far less than buying a new softswitch.
  • What is the cost of the data connection needed to connect the two locations? Generally companies make these connections using establish Internet backbone and this typically adds nothing to that cost.
  • How much does the switch owner want to charge for using the softswitch core? There are many ways to charge for this service. It could be done on a flat rate monthly lease, a connection fee per telephone number.
  • Reliability. How good is the Internet connection between the two companies? If that connection is lost then voice processes will stop. Ideally the connection ought to be on a ring or have redundant routing.
  • Control. I often hear a company thinking of leasing say that they feel like they don’t have enough control in a shared switch. But you can get access to all important switch functions remotely.

When I look at the numbers I find that it is almost always better to share a switch rather than to buy a whole new switch. However, the one factor that still often drives the decision to buy a new switch rather than share is the reliability of the data connection between the two parties. But this is generally about the same reliability as connecting remote switches in your own network to a host switch.

If you want to consider sharing a softswitch call CCG and we can help you work though this decision.

Voice – Still Relevant

A landline telephone

A landline telephone (Photo credit: Wikipedia)

At the beginning of this year the Center For Disease Control issued a report called Wireless Substitution: Early Release of Estimates From the National Health Survey, January – June 2012. The summary of that report is attached here as Wireless Substitution 2012.

The CDC asks over 20,000 households each year a number of questions associated with health issues, and starting a few years ago they started asking about basic telephone coverage. In 2012 they expanded the telephony questions to include questions about landline and cellphone usage.

The results of the study are statistically reliable. They worked hard to get the sample they use to look like America as a whole and the results are 95% accurate, plus or minus 5%. This means that if they asked everybody in the country these same questions, the results would be within 5% of the results of the survey, which is very accurate for a survey.

Some of what they found was very interesting. The most interesting statistic to me is that 65% of all households still have a landline telephone. As late as 2000 the industry had about a 98% penetration of households. As the chart on the first page of the report shows (showing wireless-only households), landline subscribers dropped slowly until about 2005 and have dropped at a steady pace since then due to migration to cell phone usage.

I have heard for years from experts who have declared the voice business dead. Many new telecom ventures are launching without a voice offering, because they believe it is irrelevant. The most visible of these is Google in Kansas City, who offers only data and cable TV. But I look at a product that is still in 65% of households as something that is still very attractive from the carrier perspective. Cable TV only has a nationwide penetration of just over 75% and even after all of the years of decline, voice is still not that far behind cable.

Voice is a very profitable business. If Google wanted to offer voice in Kansas City they would need to buy a softswitch, which might cost $1 million for a market that large. And they would have to interconnect with the existing PSTN. They would have to cover some one-time costs for each customer like number portability. After covering those costs everything else is profit. Voice can be auto-provisioned so that it doesn’t take any people to activate it. And it can be sold in a very simple package so that there are not a lot of options. Voice doesn’t need to be a complicated product.

From a business plan perspective, not offering voice is leaving a lot of low-hanging fruit on the table. For Google in Kansas City, the breakeven on paying for the voice investment could be measured in terms of a handful of months. After that it would add significantly to the margin per customer and to bottom line.

I have a hard time understanding why Google or anybody would not offer voice. With residential customers it is low-hanging fruit. And voice is still mandatory to get many business customers. Businesses bore the brunt of the competitive CLEC push a decade ago and many of them were burned by having one vendor for voice and another for data. When something went wrong both vendors would point at each other rather than fix the problem. And so a lot of businesses insist on buying all of their telecom products from a single vendor. Further, most businesses care more about reliability for voice than they do price. Voice is the lifeline of many businesses and they want it to be served by a capable vendor on a reliable network.

When Google approaches a business and wants them to buy a 1 Gigabit data pipe they are basically telling that business to keep their voice on copper or relegate it to somebody selling VoIP on the Internet. There are many companies selling VoIP this way and the quality of the connections vary widely. There are good products sold this way, but also some really sketchy stuff, so a business has to be very wary. Most businesses are just not willing to take a chance buying voice from a vendor they don’t know and who doesn’t have people in their market. They have been down that path before.

Most of my clients still offer voice services and all of them do pretty well doing so. My clients who sell to business customers report that voice is still the way to get into the door. Many of these clients are now selling IP Centrex, but that is still a voice product.

And so I look at Google and other providers who have elected to not sell voice and just scratch my head. Are they afraid of being regulated? In most states regulation of competitive voice providers is very light. Do they think voice is just obsolete and not worth the effort? This survey and all of my clients who sell voice demonstrate that this is just not the case. Voice is still very relevant and still very profitable.

Who is Going to Pay for the IP Network?

Peninsular Telephone Company

Peninsular Telephone Company (Photo credit: Nick Suan)

Small telcos and most CLECs are waiting to see what will come from the changes due to converting to an all IP network for telephony. Today the telephony voice network utilizes TDM (time division multiplexing) technology that was originally developed for copper but that has been upgraded to use fiber. But the FCC has said that this old network is going to have to be upgraded to all-IP, meaning that voice will be carried by Ethernet similar to the way that data is transmitted.

I don’t think anybody is arguing that this kind of shift makes sense. IP trunking is far more efficient in terms of carrying more calls in the same amount of bandwidth. And a lot of companies have already implemented some IP trunking.

The important issue for small telcos and CLECs is how this transition is going to change their costs. In order to understand the possible change, let’s look at how voice traffic gets to and from small telcos and CLECs today.

  • Independent telephone companies connect with larger companies and neighboring companies by physical interconnection at mutual meetpoints. Historically, most of the meetpoints are located at the physical border between two neighboring telephone companies with each company owning the fiber and electronics in their own territory. And each telco is responsible for the costs of their portion of the network. Historically local calls have been exchanged for free in both directions and there are access charges in place for all telcos to get paid by the long distance carriers for using their network and facilities for long distance calls.
  • The rules governing CLECs were established by the Telecommunications Act of 1996. This Act laid forth the basic rule that a CLEC can interconnect with a telco network at any technically feasible point. This idea was fought hard by the large telcos who wanted CLECs to bring traffic to their tandems (regional hub offices). Once a CLEC has established a meetpoint, then it works pretty much the same as normal telco interconnection in that both parties are responsible for costs on their side of the interconnection. Sometimes local calls are interchanged for a fee and sometimes they are free (called bill and keep) and this is negotiated. The CLECs also bill access charges for carrying long distance calls.

There are a number of ways that IP trunking could be implemented, and each of them has financial consequences for small telcos and CLECs:

  • The IP network could be built to mimic the current PSTN. The routes would be roughly the same but the rules of interconnection would stay the same. But with IP trunking the network would be more efficient.
  • The large telcos could establish regional hubs and expect everybody else to somehow get their traffic to those locations. This would be a radical change for small telcos who would have to build or lease fiber from their rural location to the nearest regional hub. For CLECs this would completely undo the rules established by the Telecommunications Act of 1996 and would put all of the cost to get to the hubs onto them.
  • In the most extreme IP network there would be only a few large hubs to cover the whole US. This would be the most efficient in terms of the hubs, but it would require all telcos and CLECs to spend a lot of money to get their voice traffic to and from the hub.

Since I have been working in the industry the RBOCs (now AT&T and Verizon) have tried several times to put the burden and the cost of transporting calls onto the small telcos. But regulators have always stepped in to stop this because they realize that it would greatly jack up the cost of doing business in rural areas. I certainly hope that as we move to a more efficient network that we don’t end up breaking a system that is working well.

The downside to any plan that shifts cost to small telcos is that the cost of providing local and long distance service will increase in rural areas. The consequence of changing the CLEC rules will be less competition. The current interconnection and compensation rules have served the country well. Every caller benefits by having affordable rates to call to and from rural areas. And there is no doubt that higher communications cost would be a major hindrance to creating and keeping jobs in rural areas.