 |
|
|
Telecommunications networks around the world are relying more and more on SS7 for communications between switches, databases and other network elements. While SS7 sends more signaling data per call, and at much higher speeds than older tone-based signaling methods, it is not without problems. In fact, its addressing problems are so severe, that a completely new signaling system may be required, perhaps including some lessons learned from internet protocols.
SS7 is used around the world, but there is not a global SS7 network. Rather, SS7 is a collection of national networks, with communication between national networks much more complex than within the networks. Each national SS7 network may differ in significant ways. Most importantly, the basic form of SS7 addressing ends at a country’s border, and the more complex global title addressing is required. But even then, global title translation is more complex and less available between countries as opposed to within countries.
The basic SS7 address is a Point Code, simply a number that identifies a specific switch, database or other device on an SS7 network. Point codes in the US and Canada are 24 bits long, allowing over 16 million different addresses. However, in other countries, point codes are different lengths: 14 bits in many, but also 16 bits (e.g. in Japan) or 24 bits (e.g. in China). The critical limiting factor is that point codes, no matter what their length, are re-used in each country and cannot be used for communications between national SS7 networks.
SS7 also provides global title addressing, which basically takes a phone or card identifier and uses it for routing. For example, a wireless phone may be identified by a MIN (Mobile Identification Number) or IMSI (International Mobile Station Identity). These numbers also identify the home system (HLR) of the mobile. Through global title translation, messages can be routed through the SS7 network to the home system without the originator of the message knowing the actual point code of this system. This routing is performed by STPs (Signal Transfer Points) that examine part or all of the global title and translate it into the point code of the next STP or the destination. Global titles can even be used internationally, if (and it is a big if) the international gateway is programmed to understand the specific global title, and the global title is supported by both national SS7 networks.
Compare the SS7 approach to the internet. Both have two-level addressing schemes. Everybody reading this article is familiar with the higher level internet addressing scheme (e.g. an email address like David.Crowe@cnp-wireless.com or a website address like http://www.cnp-wireless.com). These addresses contain a domain name (cnp-wireless.com) that can be converted into a TCP/IP numeric address through a domain name query. These numeric addresses are usually hidden from users, but configuring an internet connection usually requires entering at least one numeric TCP/IP address of the form xxx.xxx.xxx.xx. The internet method for routing messages is to first present the domain name to a Domain Name Server and then use numeric routing on the TCP/IP numeric address that it provides. This contrasts with the SS7 approach where global title translation may occur in a piecemeal fashion at each STP or international gateway in parallel with routing of the message. Another striking contrast is that TCP/IP provides a completely global address, allowing people around the world to access internet resources without regard to the country in which they are physically located.
These differences would be of purely academic interest if the SS7 approach did not cause any problems. However, the national scope of an SS7 point code is a potentially fatal problem, seriously impacting the development of international roaming and other new services. Wireless networks are avoiding the problem in the short term by using US point codes to extend the ANSI (US/Canadian) SS7 network into other countries to allow international roaming to occur. However, this forces countries that want to roam with the US to support two parallel SS7 networks, one running the national SS7 protocol for wireline interconnection, freephone queries, etc. and another, ANSI, network for international roaming.
Beyond the costs and political problems that extending the ANSI network internationally causes, the distributed nature of global title translation makes it extremely difficult and costly to implement each new global title, even though requirements for new global titles are occurring all the time. Every STP and international gateway in the world may have to be reconfigured (and quite possibly upgraded with new software) to support each new global title. Local Number Portability has created a further mess for global title translation by forcing several global title translations to be cloned. One global title takes a potentially ported number and routes to the LNP database, while the other takes the routing number (e.g. LRN – Location Routing Number) and routes to the destination switch or database.
An “internet” approach to addressing would make the SS7 network much more effective. First, point codes should be global, so that any network element around the world can be addressed. Secondly global title translations should be performed by the originating network element by a query to a local database, based on the type of address (e.g. directory number, credit card number, IMSI) and the type of destination required (e.g. calling name database, HLR, Short Message Center). The local database may query other databases transparently, but will eventually obtain and return the global point code of the destination. Routing can then be performed using the global point code, which can be stored to avoid subsequent queries.
It may not be possible to make such drastic changes to SS7, because the addressing concepts are implemented in such a low level in the network. However, I believe that it will become obvious to more people than myself that something needs to be done, even if a completely new protocol needs to be developed. I used to think that global title addressing was one of the coolest parts of SS7, but now I believe it is having a chilling effect on the implementation of new applications, and even on the maintenance of existing capabilities when new concepts such as number portability have to be embraced.
TCP/IP is almost certainly not the right protocol for the telephony network. It is optimized for relatively large packet data transmissions, not signaling. Using the same protocol in telephony as in the internet would also increase the danger of leaks. Just imagine the impact of ‘denial of service’ attacks by a hacker who discovered a way to send a flood of TCP/IP messages from their basement computer into the telephony network! Furthermore, although I consider some of the addressing concepts of TCP/IP to be significantly superior to SS7, the internet protocols have many of their own problems, including routing inefficiencies and potential exhaustion of their current address format.
Eventually, telecom carriers are going to have to look at a replacement for SS7 (you heard it here first). “SS8” will have to be designed with a global viewpoint or it will just be SS7 deja vu. While a new protocol will be a major expense for the telecom industry, it has the potential to stem the death-by-a-thousand-cuts that SS7 is currently causing the industry.
© – Copyright