In the cellular world, subscribers are charged for data from the IP, transport and applications layers; this means you pay for the IP header, you pay for the TCP/UDP header, and you pay for the contents (the cat videos it contains).
This also means if an operator moves mobile subscribers from IPv4 to IPv6, there’s an extra 20 bytes the customer is charged for for every packet sent / received, which the customer is charged for – This is because the IPv6 header is longer than the IPv4 header.
In most cases, mobile subs don’t get a choice as to if their connection is IPv4 or IPv6, but on a like for like basis, we can say that if a customer moves is on IPv6 every packet sent/received will have an extra 20 bytes of data consumed compared to IPv4.
This means subscribers use more data on IPv6, and this means they get charged for more data on IPv6.
For IoT applications, light users and PAYG users, this extra 20 bytes per packet could add up to something significant – But how much?
We can quantify this, but we’d need to know the number of packets sent on average, and the quantity of the data transferred, because the number of packets is the multiplier here.
So for starters I’ve left a phone on the desk, it’s registered to the network but just sitting in Idle mode – This is an engineering phone from an OEM, it’s just used for testing so doesn’t have anything loaded onto it in terms of apps, it’s not signed into any applications, or checking in the background, so I thought I’d try something more realistic.
So to get a clearer picture, I chucked a SIM in my regular everyday phone I use personally, registered it to the cellular lab I have here. For the next hour I sniffed the GTP traffic for the phone while it was sitting on my desk, not touching the phone, and here’s what I’ve got:
Overall the PCAP includes 6,417,732 bytes of data, but this includes the transport and GTP headers, meaning we can drop everything above it in our traffic calculations.
For this I’ve got 14 bytes of ethernet, 20 bytes IP, 8 bytes UDP and 5 bytes for TZSP (this is to copy the traffic from the eNB to my local machine), then we’ve got the transport from the eNB to the SGW, 14 bytes of ethernet again, 20 bytes of IP , 8 bytes of UDP and 8 bytes of GTP then the payload itself. Phew. All this means we can drop 97 bytes off every packet.
We have 16,889 packets, 6,417,732 bytes in total, minus 97 bytes from each gives us 1,638,233 of headers to drop (~1.6MB) giving us a total of 4.556 MB traffic to/from the phone itself.
This means my Android phone consumes 4.5 MB of cellular data in an hour while sitting on the desk, with 16,889 packets in/out.
Okay, now we’re getting somewhere!
So now we can answer the question, if each of these 16k packets was IPv6, rather than IPv4, we’d be adding another 20 bytes to each of them, 20 bytes x 16,889 packets gives 337,780 bytes (~0.3MB) to add to the total.
If this traffic was transferred via IPv6, rather than IPv4, we’d be looking at adding 20 bytes to each of the 16,889 packets, which would equate to 0.3MB extra, or about 7% overhead compared to IPv4.
But before you go on about what an outrage this IPv6 transport is, being charged for those extra bytes, that’s only one part of the picture.
There’s a reason operators are finally embracing IPv6, and it’s not to put an extra 7% of traffic on the network (I think if you asked most capacity planners, they’d say they want data savings, not growth).
IPv6 is, for lack of a better term, less rubbish than IPv4.
There’s a lot of drivers for IPv6, and some of these will reduce data consumption. IPv6 is actually your stuff talking directly to the remote stuff, this means that we don’t need to rely on NAT, so no need to do NAT keepalives, and opening new sessions, which is going to save you data. If you’re running apps that need to keep a connection to somewhere alive, these data savings could negate your IPv6 overhead costs.
Will these potential data savings when using IPv6 outweigh the costs?
That’s going to depend on your use case.
If you’ve extremely bandwidth / data constrained, for example, you have an IoT device on an NTN / satellite connection, that was having to Push data every X hours via IPv4 because you couldn’t pull data from it as it had no public IP, then moving it to IPv6 so you can pull the data on the public IP, on demand, will save you data. That’s a win with IPv6.
If you’re a mobile user, watching YouTube, getting push notifications and using your phone like a normal human, probably not, but if you’re using data like a normal user, you’ve probably got a sizable data allowance that you don’t end up fully consuming, and the extra 20 bytes per packet will be nothing in comparison to the data used to watch a 2k video on your small phone screen.
Ask someone with headphones and a lanyard in the halls of a datacenter what transport does DNS use, there’s a good chance the answer you’d get back is UDP Port 53.
But not always!
In scenarios where the DNS response is large (beyond 512 bytes) a DNS query will shift over to TCP for delivery.
How does the client know when to shift the request to TCP – After all, the DNS server knows how big the response is, but the client doesn’t.
The answer is the Truncated flag, in the response.
The DNS server sends back a response, but with the Truncated bit set, as per RFC 1035:
TC TrunCation – specifies that this message was truncated due to length greater than that permitted on the transmission channel.
RFC 1035
Here’s an example of the truncated bit being set in the DNS response.
The DNS client, upon receiving a response with the truncated bit set, should run the query again, this time using TCP for the transport.
One prime example of this is DNS NAPTR records used for DNS in roaming scenarios, where the response can quite often be quite large.
If it didn’t move these responses to TCP, you’d run the risk of MTU mismatches dropping DNS. In that half of my life has been spent debugging DNS issues, and the other half of my life debugging MTU issues, if I had MTU and DNS issues together, I’d be looking for a career change…
I had a question recently on LinkedIn regarding how to preference Voice over WiFi traffic so that a network engineer operating the WiFi network can ensure the best quality of experience for Voice over WiFi.
Voice over WiFi is underpinned by the ePDG – Evolved Packet Data Gateway (this is a fancy IPsec tunnel we authenticate to using the SIM to drop our traffic into the P-CSCF over an unsecured connection). To someone operating a WiFi network, the question is how do we prioritise the traffic to the ePDGs and profile it?
ePDGs can be easily discovered through a simple DNS lookup, once you know the Mobile Network Code and Mobile Country code of the operators you want to prioritise, you can find the IPs really easily.
ePDG addresses take the form epdg.epc.mncXXX.mccYYY.pub.3gppnetwork.org so let’s look at finding the IPs for each of these for the operators in a country:
The first step is nailing down the mobile network code and mobile country codes of the operators you want to target, Wikipedia is a great source for this information. Here in Australia we have the Mobile Country Code 505 and the big 3 operators all support Voice over WiFi, so let’s look at how we’d find the IPs for each. Telstra has mobile network code (MNC) 01, in 3GPP DNS we always pad network codes to 3 digits, so that’ll be 001, and the mobile country code (MCC) for Australia is 505. So to find the IPs for Telstra we’d run an nslookup for epdg.epc.mnc001.mcc505.pub.3gppnetwork.org – The list of IPs that are returned, are the IPs you’ll see Voice over WiFi traffic going to, and the IPs you should provide higher priority to:
The same rules apply in other countries, you’d just need to update the MNC/MCC to match the operators in your country, do an nslookup and prioritise those IPs.
Generally these IPs are pretty static, but there will need to be a certain level of maintenance required to keep this list up to date by rechecking.
There’s old joke about standards that the great thing about standards there’s so many to choose from.
SMS wasn’t there from the start of GSM, but within a year of the inception of 2G we had SMS, and we’ve had SMS, almost totally unchanged, ever since.
In a recent Twitter exchange, I was asked, what’s the best way to transport SMS? As always the answer is “it depends” so let’s take a look together at where we’ve come from, where we are now, and how we should move forward.
How we got Here
Between 2G and 3G SMS didn’t change at all, but the introduction of 4G (LTE) caused a bit of a rethink regarding SMS transport.
Early builders of LTE (4G) networks launched their 4G offerings without 4G Voice support (VoLTE), with the idea that networks would “fall back” to using 2G/3G for voice calls.
This meant users got fast data, but to make or receive a call they relied on falling back to the circuit switched (2G/3G) network – Hence the name Circuit Switched Fallback.
Falling back to the 2G/3G network for a call was one thing, but some smart minds realised that if a phone had to fall back to a 2G/3G network every time a subscriber sent a text (not just calls) – And keep in mind this was ~2010 when SMS traffic was crazy high; then that would put a huge amount of strain on the 2G/3G layers as subs constantly flip-flopped between them.
The SGs-AP interface has two purposes; One, It can tell a phone on 4G to fallback to 2G/3G when it’s got an incoming call, and two; it can send and receive SMS.
SMS traffic over this interface is sometimes described as SMS-over-NAS, as it’s transported over a signaling channel to the UE.
This also worked when roaming, as the MSC from the 2G/3G network was still used, so SMS delivery worked the same when roaming as if you were in the home 2G/3G network.
Enter VoLTE & IMS
Of course when VoLTE entered the scene, it also came with it’s own option for delivering SMS to users, using IP, rather than the NAS signaling. This removed the reliance on a link to a 2G/3G core (MSC) to make calls and send texts.
This was great because it allowed operators to build networks without any 2G/3G network elements and build a fully standalone LTE only network, like Jio, Rakuten, etc.
In roaming scenarios, S8 Home Routing for VoLTE enabled SMS to be handled when roaming the same way as voice calls, which made SMS roaming a doddle.
4G SMS: SMS over IP vs SMS over NAS
So if you’re operating a 4G network, should you deliver your SMS traffic using SMS-over-IP or SMS-over-NAS?
Generally, if you’ve been evolving your network over the years, you’ve got an MSC and a 2G/3G network, you still may do CSFB so you’ve probably ended up using SMS over NAS using the SGs-AP interface. This method still relies on “the old ways” to work, which is fine until a discussion starts around sunsetting the 2G/3G networks, when you’d need to move calling to VoLTE, and SMS over NAS is a bit of a mess when it comes to roaming.
Greenfield operators generally opt for SMS over IP from the start, but this has its own limitations; SMS over IP is has awful efficiency which makes it unsuitable for use with NB-IoT applications which are bandwidth constrained, support for SMS over IP is generally limited to more expensive chipsets, so the bargain basement chips used for IoT often don’t support SMS over IP either, and integration of VoLTE comes with its own set of challenges regarding VoLTE enablement.
5G enters the scene (Nsmsf_SMService)
5G rolled onto the scene with the opportunity to remove the SMS over NAS option, and rely purely on SMS over IP (IMS); forcing the industry to standardise on an option alas this did not happen.
This added another option for SMS delivery dependent on the access network used, and the Nsmsf_SMService interface does not support roaming.
Of course if you are using Voice over NR (VoNR) then like VoLTE, SMS is carried in a SIP message to the IMS, so this negates the need for the Nsmsf_SMService.
2G/3G Shutdown – Diameter to replace SGs-AP (SGd)
With the 2G/3G shutdown in the US operators who had up until this point been relying on SMS-over-NAS using the SGs-AP interface back to their MSCs were forced to make a decision on how to route SMS traffic, after the MSCs were shut down.
This landed with SMS-over-Diameter, where the 4G core (MME) communicates over Diameter with the SMSc.
This has adoption by all the US operators, but we’re not seeing it so widely deployed in the rest of the world.
State of Play
Option
Conditions
Notes
MAP
2G/3G Only
Relies on SS7 signaling and is very old Supports roaming
SGs-AP (SMS-over-NAS)
4G only relies on 2G/3G
Needs an MSC to be present in the network (generally because you have a 2G/3G network and have not deployed VoLTE) Supports limited roaming
SMS over IP (IMS)
4G / 5G
Not supported on 2G/3G networks Relies on a IMS enabled handset and network Supports roaming in all S8 Home Routed scenarios Device support limited, especially for IoT devices
Diameter SGd
4G only / 5G NSA
Only works on 4G or 5G NSA Better device support than 4G/5G Supports roaming in some scenarios
Nsmsf_SMService
5G standalone only
Only works on 5GC Doesn’t support roaming
The convoluted world of SMS delivery options
A Way Forward:
While the SMS payload hasn’t changed in the past 31 years, how it is transported has opened up a lot of potential options for operators to use, with no clear winner, while SMS revenues and traffic volumes have continued to fall.
For better or worse, the industry needs to accept that SMS over NAS is an option to use when there is no IMS, and that in order to decommission 2G/3G networks, IMS needs to be embraced, and so SMS over IP (IMS) supported in all future networks, seems like the simple logical answer to move forward.
And with that clear path forward, we add in another wildcard…
But, when you’ve only got a finite resource of bandwidth, and massive latencies to contend with, the all-IP architecture of IMS (VoLTE / VoNR) and it’s woeful inefficiency starts to really sting.
Of course there are potential workarounds here, Robust Header Correction (ROHC) can shrink this down, but it’s still going to rely on the 3 way handshake of TCP, TCP keepalive timers and IMS registrations, which in turn can starve the radio resources of the satellite link.
Even with SMS over 30 years old, we can still expect it to be a part of networks for years to come, even as WhatsApp / iMessage, etc, offer enhanced services. As to how it’s transported and the myriad of options here, I’m expecting that we’ll keep seeing a multi-transport mix long into the future.
For simple, cut-and-dried 4G/5G only network, IMS and SMS over IP makes the most sense, but for anything outside of that, you’ve got a toolbox of options for use to make a solution that best meets your needs.
Even before 5G was released, the arms race to claim the “fastest” speeds on LTE, NSA and SA networks has continued, with pretty much every operator claiming a “first” or “fastest”.
I myself have the fastest 5G network available* but I thought I’d look at how big the values are we can put in for speed, these are the Maximum Bitrate Values (like AMBR) we can set on an APN/DNN, or on a Charging Rule.
*Measurement is of the fastest 5G network in an eastward facing office, operated by a person named Nick, in a town in Australia. Other networks operated by people other than those named Nick in eastward facing office outside of Australia were not compared.
The answer for Release 8 LTE is 4294967294 bytes per second, aka 4295 Mbps 4.295 Gbps.
Not bad, but why this number?
The Max-Requested-Bandwidth-DL AVP tells the PGW the max throughput allowed in bits per second. It’s a Unsigned32 so max value is 4294967294, hence the value.
But come release 15 some bright spark thought we may in the not to distant future break this barrier, so how do we go above this?
The answer was to bolt on another AVP – the “Extended-Max-Requested-BW-DL” AVP ( 554 ) was introduced, you might think that means the max speed now becomes 2x 4.295 Gbps but that’s not quite right – The units was shifted.
This AVP isn’t measuring bits per second it’s measuring kilobits per second.
So the standard Max-Requested-Bandwidth-DL AVP gives us 4.3 Gbps, while the Extended-Max-Requested-Bandwidth gives us a 4,295 Gbps.
We add the Extended-Max-Requested-Bandwidth AVP (4295 Gbps) onto the Max-Requested Bandwidth AVP (4.3 Gbps) giving us a total of 4,4299.3 Gbps.
I started off just updating the SPN, OPN, etc, etc, but I had a suspicion there were still references.
I confirmed this pretty easily with Wireshark, first I started a trace in Wireshark of the APDUs: I enabled capturing on a USB Interface:
modprobe usbmon
Then we need to find where our card reader is connected, running ‘lsusb‘ lists all the USB devices, and you can see here’s mine on Bus 1, Device 49.
Then fired up Wireshark, selected USB Bus 01 to capture all the USB traffic on the bus.
Then I ran the “export” command in PySIM to read the contents of all the files on the SIM, and jumped back over to Wireshark. (PySIM decodes most files but not all – Whereas this method just looks for the bytes containing the string)
From the search menu in Wireshark I searched the packet bytes for the string containing the old brand name, and found two more EFs I’d missed.
For anyone playing along at home, using this method I found references to the old brand name in SMSP (which contains the network name) and ADN (Which had the customer support number as a contact with the old brand name).
Short one, The other day I needed to add a Network Appearance on an SS7/SS7 M3UA linkset.
Network Appearances on M3UA links are kinda like a port number, in that they allow you to distinguish traffic to the same point code, but handled by different logical entities.
When I added the NA parameter on the Linkset nothing happened.
If you’re facing the same you’ll need to set:
cs7 multi-instance
In the global config (this is the part I missed).
Then select the M3UA linkset you want to change and add the network-appearance parameter:
network-appearance 10
And bingo, you’ll start seeing it in your M3UA traffic:
Last year I purchased a cheap second hand Huawei macro base station – there’s lots of these on the market at the moment due to the fact they’re being replaced in many countries.
I’m using it in my lab environment, and as such the config I’ve got is very “bare bones” and basic. Keep in mind if you’re looking to deploy a Macro eNodeB in production, you may need more than just a blog post to get everything tuned and functioning properly…
In this post we’ll cover setting up a Huawei BTS3900 eNodeB from scratch, using the MML interface, without relying on the U2020 management tool.
Obviously the details I setup (IP Addressing, PLMN and RF parameters) are going to be different to what you’re configuring, so keep that in mind, where I’ve got my MME Addresses, site IDs, TACs, IP Addresses, RFUs, etc, you’ll need to substitute your own values.
A word on Cabinets
Typically these eNodeBs are shipped in cabinets, that contain the power supplies, alarm / environmental monitoring, power distribution, etc.
Early on in the setup process we’ll be setting the cabinet types we’ve got, and then later on we’ll tell the system what we have installed in which slots.
This is fine if you have a cabinet and know the type, but in my case at least I don’t have a cabinet manufactured by Huawei, just a rack with some kit mounted in it.
This is OK, but it leads to a few gotchas I need to add a cabinet (even though it doesn’t physically exist) and when I setup my RRUs I need to define what cabinet, slot and subrack it’s in, even though it isn’t in any. Keep this in mind as we go along and define the position of the equipment, that if you’re not using a real-world cabinet, the values mean nothing, but need to be kept consistent.
To begin we’ll need to setup the basics, by disabling DHCP and setting an local IP Address for the unit.
SET DHCPSW: SWITCH=DISABLE;
SET LOCALIP: IP="192.168.5.234", MASK="255.255.248.0";
Obviously your IP address details will be different. Next we’ll add an eNodeB function, the LMPT / UMPT can have multiple functions and multiple eNodeBs hosted on the same hardware, but in our case we’re just going to configure one:
Again, your eNodeB ID, location, site name, etc, are all going to be different, as will your location.
Next we’ll set the system to maintenance mode (MNTMODE), so we can make changes on the fly (this takes the eNB off the air, but we’re already off the air), you’ll need to adjust the start and end times to reflect the current time for the start time, and end time to be after you’re done setting all this up.
SET MNTMODE: MNTMode=INSTALL, ST=2013&09&20&15&00&00, ET=2013&09&25&15&00&00, MMSetRemark="NewSite Install";
Next we’ll set the operator details, this is the PLMN of the eNodeB, and create a new tracking area.
Next we’ll be setting and populating the cabinets I mentioned earlier. I’ll be telling the unit it’s inside a APM30 (Cabinet 0), and in Cabinet Number 0, Subrack 0, is a BBU3900.
//To modify the cabinet type, run the following command: ADD CABINET:CN=0,TYPE=APM30; //Add a BBU3900 subrack, run the following command: ADD SUBRACK:CN=0,SRN=0,TYPE=BBU3900; //To configure boards and RF datas, run the following commands:
And inside the BBU3900 there’s some cards of course, and each card has as slot, as per the drawing below.
In my environment I’ve got a LMPT in slot 7, and a LBBP in Slot 3. There’s a fan and a UPEU too, so: We’ll add a board in Slot No. 7, of type LMPT, We’ll add a board in Slot No. 3, of type LBBP working on FDD, We’ll add a fan board in Slot No. 16, and a UPEU in Slot No. 18.
Huawei publish design guides for which cards should be in which slots, the general rule is that your LMPT / UMPT card goes in Slot 7, with your BBP cards (UBBP or LBBP) in slots 3, then 2, then 1, then 0. Fans and UPEUs can only go in the slots designed to fit them, so that makes it a bit easier.
Next we’ll need to setup our RRUs, for this we’ll need to setup an RRU chain, which is the Huawei term for the CPRI links and add an RRU into it:
//Modify the reference signal power.
MOD PDSCHCFG: LocalCellId=1, ReferenceSignalPwr=-81;
//Add an operator for the cell.
ADD CELLOP: LocalCellId=0, TrackingAreaId=0;
//Activate the cell.
ACT CELL: LocalCellId=1;
The Binding Support Function is used in 4G and 5G networks to allow applications to authenticate against the network, it’s what we use to authenticate for XCAP and for an Entitlement Server.
Rather irritatingly, there are two BSF addresses in use:
If the ISIM is used for bootstrapping the FQDN to use is:
bsf.ims.mncXXX.mccYYY.pub.3gppnetwork.org
But if the USIM is used for bootstrapping the FQDN is
bsf.mncXXX.mccYYY.pub.3gppnetwork.org
You can override this by setting the 6FDA EF_GBANL (GBA NAF List) on the USIM or equivalent on the ISIM, however not all devices honour this from my testing.
One of the hyped benefits of a 5G Core Networks is that 5GC can be used for wired networks (think DSL or GPON) – In marketing terms this is called “Wireless Wireline Convergence” (5G WWC) meaning DSL operators, cable operators and fibre network operators can all get in on this sweet 5GC action and use this sexy 5G Core Network tech.
This is something that’s in the standards, and that the big kit vendors are pushing heavily in their marketing materials. But will it take off? And should operators of wireline networks (fixed networks) be looking to embrace 5GC?
Comparing 5GC with current wireline network technologies isn’t comparing apples to apples, it’s apples to oranges, and they’re different fruits.
At its heart, the 3GPP Core Networks (including 5G Core) address one particular use cases of the cellular industry: Subscriber mobility – Allowing a customer to move around the network, being served by different kit (gNodeBs) while keeping the same IP Address.
The most important function of 5GC is subscriber mobility.
This is achieved through the use of encapsulating all the subscriber’s IP data into a GTP (A protocol that’s been around since 2G first added data).
Do I need a 5GC for my Fixed Network?
Wireline networks are fixed. Subscribers don’t constantly move around the network. A GPON customer doesn’t need to move their OLT every 30 minutes to a new location.
Encapsulating a fixed subscriber’s traffic in GTP adds significant processing overhead, for almost no gain – The needs of a wireline network operator, are vastly different to the needs of a cellular core.
Today, you can take a /24 IPv4 block, route it to a DSLAM, OLT or CMTS, and give an IP to 254 customers – No cellular core needed, just a router and your access device and you’re done, and this has been possible for decades. Because there’s no mobility the GTP encapsulation that is the bedrock for cellular, is not needed.
Rather than routing directly to Access Network kit, most fixed operators deploy BRAS systems used for fixed access. Like the cellular packet core, BRAS has been around for a very long time, with a massive install base and a sea of engineering experience in house, it meets the needs of the wireline industry who define its functions and roles along with kit vendors of wireline kit; the fixed industry working groups defined the BRAS in the same way the 3GPP and cellular industry working groups defined 5G Core.
I don’t forsee that we’ll see large scale replacement of BRAS by 5GC, for the same reason a wireless operator won’t replace their mobile core with a BRAS and PPPoE – They’re designed to meet different needs.
All the other features that have been added to the 3GPP Core Network functionality, like limiting speed, guaranteed throughput bearers, 5QI / QCI values, etc, are addons – nice-to-haves. All of these capabilities could be implemented in wireline networks today – if the business case and customer demand was there.
But what about slicing?
With dropping ARPUs across the board, additional services relating to QoS (“Network Slicing”) are being held up as the saving grace of revenues for cellular operators and 5G as a whole, however this has yet to be realized and early indications suggest this is not going to be anywhere near as lucrative as previously hoped.
What about cost savings?
In terms of cost-per-bit of throughput, the existing install base wireline operators have of heavy-metal kit capable of terabit switching and routing has been around for some time in fixed world, and is what most 5G Cores will connect to as their upstream anyway, so there won’t be any significant savings on equipment, power consumption or footprint to be gained.
Fixed networks transport the majority of the world’s data today – Wireline access still accounts for the majority of traffic volumes, so wireline kit handles a higher magnitude of throughput than it’s Packet Core / 5GC cousins already.
Cutting down the number of parts in the network is good though right?
If you’re operating both a Packet Core for Cellular, and a fixed network today, then you might think if you moved from the traditional BRAS architecture fore the wired network to 5GC, you could drop all those pesky routers and switches clogging up your CO, Exchanges and Data Centers.
The problem is that you still need all of those after the 5GC to be able to get the traffic anywhere users want to go. So the 5GC will still need all of that kit, all your border routers and peering routers will remain unchanged, as well as domestic transmission, MPLS and transport.
The parts required for operating fixed networks is actually pretty darn small in comparison to that of 5GC.
TL;DR?
While cellular vendors would love to sell their 5GC platform into fixed operators, the premise that they are willing to replace existing BRAS architectures with 5GC, is as unlikely in my view as 5GC being replaced by BRAS.
We recently added support in PyHSS for fixed line SIP subscribers to attach to the IMS.
Traditional telecom operators are finding their fixed line network to be a bit of a money pit, something they’re required to keep operating to meet regulatory obligations, but the switches are sitting idle 99% of the time. As such we’re seeing more and more operators move fixed line subs onto their IMS.
This new feature means we can use PyHSS to serve as the brains for a fixed network, as well as for mobile, but there’s one catch – How we authenticate subscribers changes.
Most banks of line cards in a legacy telecom switches, or IP Phones, don’t have SIM slots to allow us to authenticate, so instead we’re forced to fallback to what they do support.
Unfortunately for the most part, what is supported by these IP phones or telecom switches is SIP MD5 Digest Authentication.
The Nonce is generated by the HSS and put into the Multimedia-Authentication-Answer, along with the subscriber’s password and sent in the clear to the S-CSCF.
The HSS then generates the the Multimedia-Auth Answer, it generates a nonce (in the 3GPP-SIP-Authenticate / 609 AVP) and sends the Subscriber’s password in the 3GPP-SIP-Authorization (610) AVP in response back to the S-CSCF.
I would have thought a better option would be for the HSS to generate the Nonce and Digest, and then the S-CSCF to just send the Nonce to the Sub and compare the returned Digest from the Sub against the expected Digest from the HSS, but it would limit flexibility (realm adaptation, etc) I guess.
The UE/UA (I guess it’s a UA in this context as it’s not a mobile) then generates its own Digest from the Nonce and sends it back to the S-CSCF via the P-CSCF.
The S-CSCF compares the received Digest response against the one it generated, and if the two match, the sub is authenticated and allowed to attach onto the network.
In the past I had my iFCs setup to look for the P-Access-Network-Info header to know if the call was coming from the IMS, but it wasn’t foolproof – Fixed line IMS subs didn’t have this header.
The first thing people learn about SIMs or the Smart Cards that the SIM / USIM app runs on, is that “There’s a little computer in the card”. So how little is this computer, and what’s the computing power in my draw full of SIMs?
So for starters the SIM manufacturers love their NDAs, so I can’t post the chip specifications for the actual cards in my draw, but here’s some comparable specs from a seller selling Java based smart cards online:
Specs for Smart Card
4K of RAM is 4069 bytes. For comparison the Apollo Guidance Computer had 2048 words of RAM, but each “word” was 16 bits (two bytes), so actually this would translate to 4069 bytes so equal with one of these smart cards in terms of RAM – So the smart card above is on par with the AGC that took humans to the moon in terms of RAM, althhough the SIMs would be a wee bit larger if they were also using magnetic core memory like the AGC!
The Nintendo Entertainment System was powered by a MOS Technology 6502, it had access to 2K of RAM, two the Smart Card has twice as much RAM as the NES, so it could get you to the moon and play Super Mario Bros.
What about comparing Non-Volatile Memory (Storage)? Well, the smart card has 145KB of ROM / NVM, while Apollo flew with 36,864 words of RAM, each word is two bits to 73,728 Bytes, so roughly half of what the Smart Card has – Winner – Smart Card, again, without relying on core rope memory like AGC.
SIM cards are clocked kinda funkily so comparing processor speeds is tricky. Smart Cards are clocked off the device they connect to, which feeds them a clock signal via the CLK pin. The minimum clock speed is 1Mhz while the max is 5Mhz.
Now I’m somewhat of a hoarder when it comes to SIM Cards; in the course of my work I have to deal with a lot of SIMs…
Generally when we’re getting SIMs manufactured, during the Batch Approval Process (BAP) the SIM vendor will send ~25 cards for validation and testing. It’s not uncommon to go through several revisions. I probably do 10 of these a year for customers, so that’s 250 cards right there.
Then when the BAP is done I’ll get another 100 or so production cards for the lab, device testing, etc, this probably happens 3 times a year.
So that’s 550 SIMs a year, I do clean out every so often, but let’s call it 1000 cards in the lab in total.
In terms of ROM that gives me a combined 141.25 MB, I could store two Nintendo 64 games, or one Mini CD of data, stored across a thousand SIM cards – And you thought installing software from a few floppies was a pain in the backside, imagine accessing data from 1000 Smart Cards!
What about tying the smart cards together to use as a giant RAM BUS? Well our 1000 cards give us a combined 3.91 MB of RAM, well that’d almost be enough to run Windows 95, and enough to comfortably run Windows 3.1.
Practical do do any of this? Not at all, now if you’ll excuse me I think it’s time I throw out some SIMs…
If you’ve ever worked in roaming, you’ll probably have had the misfortune of dealing with Transferred Account Procedures aka TAP files.
It’s used for billing a 2G GSM call right up to 5G data usage, if you use a service while roaming, somewhere in the world there’s a TAP file with your usage in it.
A brief history of TAP
TAP was originally specified by the GSMA in 1991 as a standard CDR interchange format between operators, for use in roaming scenarios.
Notice I said GSMA – Not 3GPP – This means there’s no 3GPP TS docs for this, it’s defined by the industry lobby group’s members, rather than the standards body.
So what does this actually mean? Well, if you’re MNO A and a customer from MNO B roams into your network, all the calls, SMS and data consumed by the roaming subscriber from MNO B will need to be billed to MNO B, by you, MNO A.
If a network operator wants to get paid for traffic used on their network by roaming subscribers, they’d better send out a TAP file to the roamer’s home network.
TAP is the file format generated my MNO A and sent to MNO B, containing all the usage charges that subscribers from MNO B have racked up while roaming into your network.
These are broken down into “Transactions” (CDRs), for events like making a call, connecting a PDN session and consuming data, or sending a text.
In the beginning of time, GSM provided only voice calling service. This meant that the only services a subscriber could consume while roaming was just making/receiving voice calls which were billed at the end of each month. – This meant billing was equally simple, every so often the visisted network would send the TAP files for the voice calls made by subscribers visited other networks, to the home networks, which would markup those charges, and add them onto the monthly invoice for each subscriber who was roaming.
But of course today, calling accounts for a tiny amount of usage on the network, but this happened gradually while passing through the introduction of SMS, CAMEL services, prepaid services, mobile data, etc. For all these services that could be offered, the TAP format had to evolve to handle each of these scenarios.
As we move towards a flat IP architecture, where voice calls and SMS sent while roaming are just data, TAP files for 4G and 5G networks only need to show data transactions, so the call objects, CAMEL parameters and SMS objects are all falling by the wayside.
What’s inside a TAP File
TAP uses the most beloved of formats – ASN1 to encode the data. This means it is strictly formatted and rigidly specified.
Each file contains a Sequence Number which is a monotonically increasing number, which allows the receiver to know if any files have been missed between the file that’s being currently parsed, an the previous file.
They also have a recipient and sender TADIG code, which is a code allocated by GSMA that uniquely identifies the sender and the recipient of the file.
The TAP records exist in one of two common format, Notification Records and transferBatch records.
These files are exchanged between operators, in practice this means “Dumped on an FTP server as agreed between the two”.
TAP Notification Records
Notifications are the simplest of TAP records and are used when there aren’t any CDRs for roaming events during the time period the TAP file covers.
These are essentially blank TAP files generated by the visited network to let the home network know it’s still there, but there are no roaming subs consuming services in that period.
Notification files are really simple, let’s take a look as one shown as JSON:
When we have services to bill and records to charge, that’s when instead we generate a transferBatch record.
It looks something like this:
There’s a lot going on in here, so let’s break it down section by section.
accountingInfo
The accountingInfo section specifies the currency, exchange rate parameters.
Keep in mind a TAP record generated by an operator in the US, would use USD, while the receiver of the file may be a European MNO dealing in EUR.
This gets even more complicated if you’re dealing with more obscure currencies where an intermediary currency is used, that’s where we bring in SDRs (“Special Drawing Right”) that map to the dollar value to be charged, kinda – the roaming agreement defines how many SDRs are in a dollar, in the example below we’re not using any, but you do see it.
When it comes to numbers and decimal places, TAP doesn’t exactly make it easy.
Significant Digits are defined by counting the first number before the decimal point and all the numbers to the right of the decimal point, so for example the number 1.234 would be 4 significant digits (1 digit before the decimal point and 3 digits after it).
Decimal Places are not actually supported for the Value fields in the TAP file. This is tricky because especially today when roaming tariffs are quite low, these values can be quite small, and we need to represent them as an integer number. TAP defines decimal places as the number of digits after the decimal place.
When it comes to the maximum number of decimal places, this actually impacts the maximum number we can store in the field – as ASN1 strictly enforce what we put in it.
The auditControlInfo section contains the number of CDRs (callEventDetailsCount) contained in the TAP file, the timestamp of the first and last CDR in the file, the total charge and any tax charged.
All of the currency information was provided in the accountingInfo so this is just giving us our totals.
A CDR has 30 days from the time it was generated / service consumed by the roamer, to be baked into a TAP file. After this we can no longer charge for it, so it’s important that the earliestCallTimeStamp is not more than 30 days before the fileCreationTimeStamp seen in batchControlInfo.
batchControlInfo
The batchControlInfo section specifies the time the TAP file became available for transfer, the time the file was created (usually the same), the sequence number and the sender / recipient TADIG codes.
As mentioned earlier, we track sequence number so the receiver can know if a TAP file has been missed; for example if you’ve got TAP file 1 and TAP file 3 comes in, you can determine you’ve missed TAP file 2.
Now we’re getting to the meat & potatoes of our TAP record, the CDRs themselves.
In LTE networks these are just records of data consumption, so let’s take a look inside the gprsCall records under callEventDetails:
In the gprsBasicCallInformation we’ve got as the name suggests the basic info about the data usage event. The time when the session started, the charging ID, the IMSI and the MSISDN of the subscriber to charge, along with their IP and the APN used.
Next up we have the gprsLocationInformation – rates and tariffs may be set based on the location of the subscriber, so we need to identify the area the sub was using the services to select correct tariff / rate for traffic in this destination.
The recEntity is the index number of the SGW / PGW used for the transaction (more on that later).
Next we have the gprsServiceUsed which, again as the name suggests, details the services used and the charge.
chargeDetailList contains the charged data (Made up of dataVolumeIncoming + dataVolumeOutgoing) and the cost.
The chargeableUnits indicates the actual data consumed, however most roaming agreements will standardise on some level of rounding, for example rounding up to the nearest Kilobyte (1024 bytes), so while a sub may consume 1025 bytes of data, they’d be billed for 2045 bytes of data. The data consumed is indicated in the chargeableUnits which indicates how much data was actually consumed, before any rounding policies where applied, while the amount that is actually charged (When taking into account rounding policies) isindicated inside Charged Units.
In the example below data usage is rounded up to the nearest 1024 bytes, 134390 bytes rounds up to the nearest 1024 gives you 135168 bytes.
As this is data we’re talking bytes, but not all bytes are created equal!
VoLTE traffic, using a QCI1 bearer is more valuable than QCI 9 cat videos, and TAP records take this into account in the Call Type Groups, each of which has a different price – Call Type Level 1 indicates the type of traffic, for S8 Home Routed LTE Traffic this is 10 (HGGSN/HP-GW), while Call Type Level 2 indicates the type of traffic as mapped to QCI values:
So Call Type Level 2 set to 20 indicates that this is “20 Unspecified/default LTE QCIs”, and Call Type Level 3 can be set to any value based on a defined inter-operator tariff.
recEntityType 7 means a PGW and contains the IP of the PGW in the Home PLMN, while recEntityType 8 means SGW and is the SGW in the Visited PLMN.
So this means if we reference recEntityCode 2 in a gprsCall, that we’re referring to an SGW at 1.2.3.5.
Lastly also got the utcTimeOffsetInfo to indicate the timezones used and assign a unique code to it.
Using the Records
We as humans? These records aren’t meant for us.
They’re designed to be generated by the Visited PLMN and sent to to the home PLMN, which ingests it and pays the amount specified in the time agreed.
Generally this is an FTP server that the TAP records get dumped into, and an automated bank transfer job based on the totals for the TAP records.
Testing of the TAP records is called “TADIG Testing” and it’s something we’ll go into another day, but in essence it’s validating that the output and contents of the files meet what both operators think is the contract pricing and specifications.
So that’s it! That’s what’s in a TAP record, what it does and how we use it!
GSMA are introducing BCE – Billing & Charging Evolution, a new standard, designed to last for the next 30+ years like TAP has. It’s still in its early days, but that’s the direction the GSMA has indicated it would like to go.
For the past few months I’ve had a Band 78 NR active antenna unit sitting next to my desk.
It’s a very cool bit of kit that doesn’t get enough love, but I thought I’d pop open the radome and take a peek inside.
Individual antenna elements
What I found very interesting is that it’s not all antennas in there!
… 29, 30, 31, 32. Yup. Checks out.
There are the expected number of antennas (I mean if I opened it up and found 31 antennas I’d have been surprised) but they don’t take up the whole volume of the unit, only about half,
AAU with Radome reinstalled
Well, after that strip show, back to sitting in my office until I need to test something 5G SA again…
Everything was working on the IMS, then I go to bed, the next morning I fire up the test device and it just won’t authenticate to the IMS – The S-CSCF generated a 401 in response to the REGISTER, but the next REGISTER wouldn’t pass.
When we generate the vectors (for IMS auth and standard auth) one of the inputs to generate the vectors is the Sequence Number or SQN.
This SQN ticks over like an odometer for the number of times the SIM / HSS authentication process has been performed.
There is some leeway in the SQN – It may not always match between the SIM and the HSS and that’s to be expected. When the MME sends an Authentication-Information-Request it can ask for multiple vectors so it’s got some in reserve for the next time the subscriber attaches, and that’s allowed.
But there are limits to how far out our SQN can be, and for good reason – One of the key purposes for the SQN is to protect against replay attacks, where the same vector is replayed to the UE. So the SQN on the HSS can be ahead of the SIM (within reason), but it can’t be behind – Odometers don’t go backwards.
So the issue was with the SQN on the SIM being out of Sync with the SQN in the IMS, how do we know this is the case, and how do we fix this?
Well there is a resync mechanism so the SIM can securely tell the HSS what the current SQN it is using, so the HSS can update it’s SQN.
When verifying the AUTN, the client may detect that the sequence numbers between the client and the server have fallen out of sync. In this case, the client produces a synchronization parameter AUTS, using the shared secret K and the client sequence number SQN. The AUTS parameter is delivered to the network in the authentication response, and the authentication can be tried again based on authentication vectors generated with the synchronized sequence number.
In our example we can tell the sub is out of sync as in our Multimedia Authentication Request we see the SIP-Authorization AVP, which contains the AUTS (client synchronization parameter) which the SIM generated and the UE sent back to the S-CSCF. Our HSS can use the AUTS value to determine the correct SQN.
SIP-Authorization AVP in the Multimedia Authentication Request means the SQN is out of Sync and this AVP contains the RAND and AUTN required to Resync
Note: The SIP-Authorization AVP actually contains both the RAND and the AUTN concatenated together, so in the above example the first 32 bytes are the AUTN value, and the last 32 bytes are the RAND value.
So the HSS gets the AUTS and from it is able to calculate the correct SQN to use.
Then the HSS just generates a new Multimedia Authentication Answer with a new vector using the correct SQN, sends it back to the IMS and presto, the UE can respond to the challenge normally.
We’re doing more and more network automation, and something that came up as valuable to us would be to have all the IPs in HOMER SIP Capture come up as the hostnames of the VM running the service.
Luckily for us HOMER has an API for this ready to roll, and best of all, it’s Swagger based and easily documented (awesome!).
(Probably through my own failure to properly RTFM) I was struggling to work out the correct (current) way to Authenticate against the API service using a username and password.
Because the HOMER team are awesome however, the web UI for HOMER, is just an API client.
This means to look at how to log into the API, I just needed to fire up Wireshark, log into the Web UI via my browser and then flick through the packets for a real world example of how to do this.
Homer Login JSON body as seen by Wireshark
In the Login action I could see the browser posts a JSON body with the username and password to /api/v3/auth
And in return the Homer API Server responds with a 201 Created an a auth token back:
Now in order to use the API we just need to include that token in our Authorization: header then we can hit all the API endpoints we want!
For me, the goal we were setting out to achieve was to setup the aliases from our automatically populated list of hosts. So using the info above I setup a simple Python script with Requests to achieve this:
import requests
s = requests.Session()
#Login and get Token
url = 'http://homer:9080/api/v3/auth'
json_data = {"username":"admin","password":"sipcapture"}
x = s.post(url, json = json_data)
print(x.content)
token = x.json()['token']
print("Token is: " + str(token))
#Add new Alias
alias_json = {
"alias": "Blog Example",
"captureID": "0",
"id": 0,
"ip": "1.2.3.4",
"mask": 32,
"port": 5060,
"status": True
}
x = s.post('http://homer:9080/api/v3/alias', json = alias_json, headers={'Authorization': 'Bearer ' + token})
print(x.status_code)
print(x.content)
#Print all Aliases
x = s.get('http://homer:9080/api/v3/alias', headers={'Authorization': 'Bearer ' + token})
print(x.json())
And bingo we’re done, a new alias defined.
We wrapped this up in a for loop for each of the hosts / subnets we use and hooked it into our build system and away we go!
With the Homer API the world is your oyster in terms of functionality, all the features of the Web UI are exposed on the API as the Web UI just uses the API (something I wish was more common!).
Using the Swagger based API docs you can see examples of how to achieve everything you need to, and if you ever get stuck, just fire up Wireshark and do it in the Homer WebUI for an example of how the bodies should look.
Misunderstood, under appreciated and more capable than people give it credit for, is our PCRF.
But what does it do?
Most folks describe the PCRF in hand wavy-terms – “it does policy and charging” is the answer you’ll get, but that doesn’t really tell you anything.
So let’s answer it in a way that hopefully makes some practical sense, starting with the acronym “PCRF” itself, it stands for Policy and Charging Rules Function, which is kind of two functions, one for policy and one for rules, so let’s take a look at both.
Policy
In cellular world, as in law, policy is the rules.
For us some examples of policy could be a “fair use policy” to limit customer usage to acceptable levels, but it can also be promotional packages, services like “free Spotify” packages, “Voice call priority” or “unmetered access to Nick’s Blog and maximum priority” packages, can be offered to customers.
All of these are examples of policy, and to make them work we need to target which subscribers and traffic we want to apply the policy to, and then apply the policy.
Charging Rules
Charging Rules are where the policy actually gets applied and the magic happens.
It’s where we take our policy and turn it into actionable stuff for the cellular world.
Let’s take an example of “unmetered access to Nick’s Blog and maximum priority” as something we want to offer in all our cellular plans, to provide access that doesn’t come out of your regular usage, as well as provide QCI 5 (Highest non dedicated QoS) to this traffic.
To achieve this we need to do 3 things:
Profile the traffic going to this website (so we capture this traffic and not regular other internet traffic)
Charge it differently – So it’s not coming from the subscriber’s regular balance
Up the QoS (QCI) on this traffic to ensure it’s high priority compared to the other traffic on the network
So how do we do that?
Profiling Traffic
So the first step we need to take in providing free access to this website is to filter out traffic to this website, from the traffic not going to this website.
Let’s imagine that this website is hosted on a single machine with the IP 1.2.3.4, and it serves traffic on TCP port 443. This is where IPFilterRules (aka TFTs or “Traffic Flow Templates”) and the Flow-Description AVP come into play. We’ve covered this in the past here, but let’s recap:
IPFilterRules are defined in the Diameter Base Protocol (IETF RFC 6733), where we can learn the basics of encoding them,
They take the format:
action dir proto from src to dst
The action is fairly simple, for all our Dedicated Bearer needs, and the Flow-Description AVP, the action is going to be permit. We’re not blocking here.
The direction (dir) in our case is either in or out, from the perspective of the UE.
Next up is the protocol number (proto), as defined by IANA, but chances are you’ll be using 17 (UDP) or 6 (TCP).
The from value is followed by an IP address with an optional subnet mask in CIDR format, for example from 10.45.0.0/16would match everything in the 10.45.0.0/16 network.
Following from you can also specify the port you want the rule to apply to, or, a range of ports.
Like the from, the tois encoded in the same way, with either a single IP, or a subnet, and optional ports specified.
And that’s it!
So let’s create a rule that matches all traffic to our website hosted on 1.2.3.4 TCP port 443,
permit out 6 from 1.2.3.4 443 to any 1-65535
permit out 6 from any 1-65535 to 1.2.3.4 443
All this info gets put into the Flow-Information AVPs:
With the above, any traffic going to/from 1.23.4 on port 443, will match this rule (unless there’s another rule with a higher precedence value).
Charging Actions
So with our traffic profiled, the next question is what actions are we going to take, well there’s two, we’re going to provide unmetered access to the profiled traffic, and we’re going to use QCI 4 for the traffic (because you’ll need a guaranteed bit rate bearer to access!).
Charging-Group for Profiled Traffic
To allow for Zero Rating for traffic matching this rule, we’ll need to use a different Rating Group.
Let’s imagine our default rating group for data is 10000, then any normal traffic going to the OCS will use rating group 10000, and the OCS will apply the specific rates and policies based on that.
Rating Groups are defined in the OCS, and dictate what rates get applied to what Rating Groups.
For us, our default rating group will be charged at the normal rates, but we can define a rating group value of 4000, and set the OCS to provide unlimited traffic to any Credit-Control-Requests that come in with Rating Group 4000.
This is how operators provide services like “Unlimited Facebook” for example, a Charging Rule matches the traffic to Facebook based on TFTs, and then the Rating Group is set differently to the default rating group, and the OCS just allows all traffic on that rating group, regardless of how much is consumed.
Inside our Charging-Rule-Definition, we populate the Rating-Group AVP to define what Rating Group we’re going to use.
Setting QoS for Profiled Traffic
The QoS Description AVP defines which QoS parameters (QCI / ARP / Guaranteed & Maximum Bandwidth) should be applied to the traffic that matches the rules we just defined.
As mentioned at the start, we’ll use QCI 4 for this traffic, and allocate MBR/GBR values for this traffic.
Putting it Together – The Charging Rule
So with our TFTs defined to match the traffic, our Rating Group to charge the traffic and our QoS to apply to the traffic, we’re ready to put the whole thing together.
So here it is, our “Free NVN” rule:
I’ve attached a PCAP of the flow to this post.
In our next post we’ll talk about how the PGW handles the installation of this rule.
