Service Level Agreements for SMS Services, pointless?

I attended a presentation about the need for SLAs in the SMS industry by Michael Kowalzik, CEO of Tyn-Tec, at Global Messaging 2008 last week.

I wasn't going to go as I had seen him speak on the same subject last year, but it was on a track I was interested in. Thankfully he spent less time this year trying to promote Tyn-Tec's services and more on what they were and why they were needed.

The 2 main areas for SMS services that differ from standard IT services are

• Message Throughput
• Time to first delivery attempt

Message throughput is fairly easy to measure and report on but time to first delivery attempt is trickier for many of us providers as it relies on the network operators passing the information back.

Tyn-Tec make a great play of running their own SMSC infrastructure, hosted on Manx Telecom's network in the Isle of Man. They have complete control and thus can get the interim delivery receipts required out of their own SMSC.

We don't run our own SMSC but connect into network operators around the world to perform the delivery on our behalf. One of my periodic bang my head against the desk tasks has been to try and get our suppliers to give us the information so we can measure their performance, report that to customers and offer SLAs on the whole delivery process not just until the message leaves our system.

That finally looks like it's bearing fruit, the operators are realising that businesses are using SMS to drive business processes and that SLAs are a minimum requirement.

Great you'd think. Possibly.

I can't help thinking that operators implementing SMS Home Routing (see  SMS Home Routing, should we as an industry be worried?) is going to invalidate the SLA in the eyes of the customer.

Giving an SLA on first attempt is all very well but the customer expects that attempt to be to the handset rather than an interim system on the destination network. The value of the SLA is immediately watered-down when you have to introduce backside covering caveats about the SLA not representing handset delivery to certain destinations.

I concluded in my previous post that SMS Home Routing was an inevitability that we has an industry had to adapt to. I wonder if its corruption of SLAs will end up representing an opportunity lost.

SMS Home Routing, should we as an industry be worried?

SMS Home Routing was a key topic discussed at Global Messaging 2008 this week. It's introduction brings with it the potential to seriously disrupt how we in the mobile messaging business deliver our services, especially cross-border. I made it my mission this week to try and understand what it means from the operator and vendor perspective

Firstly, I should probably explain the problem. SMS Home Routing means that whenever a message is sent to me it always passes through my home network, irrespective of where I am in the world. Normally when roaming, my home network never sees the message as it sent directly from the originating network to the network I'm roaming on.

Great, my operator can give me all sorts of value add services, like O2's Blue Book,SPAM filtering, twinning, out of office replies, SMS forwarding, etc, etc. To do all of these, the operator need to be able to intercept the SMS and apply some rules on it before passing it on to my handset.

It's the implementation of this interception that is the cause for concern. I spoke to, and heard from, senior figures at vendors like Airwide, Telsis and Tekelec who presented opposing views on how they recommend that their customers, the operators, do this. Similarly when I spoke to the operators they were also divided.

In the blue corner we have the transactional advocates. They believe that the transactional nature of SMS should be preserved. When a request is made to send a message to a handset, their systems proxy the request in real time. It handles delivering the message onto the handset and returns the ACK (delivery indication) back if the handset accepts the message. The sending SMSC know's nothing other of what happens behind the proxy but when the home network router reports the messages as delivered, it is.

In the red corner we have the store and forward advocates. Their systems will store the message and return the ACK back to the sending SMSC before it attempts to deliver the message to the destination handset. This unfortunately breaks the transactional nature of SMS and one of the key features that certainly most of our customers buy into, knowing whether the message has got to the handset.

I'll be honest, I arrived at the conference passionately backing the blue corner, a delivery receipt should tell me when the messages has actually arrived at the handset, but one of the vendors I spoke to certainly gave me some use cases that cause me to question this stance.

• Security - send me an SMS and you can find out which country I'm in, what if I don't know you or want you to know where I am.
• Status - a delivery receipt implies something about my status, I'm accepting messages therefore you can accept a response, what if I don't want you to know.
• Twinning - where an SMS is sent to two or more of my handsets simultaneously when is it delivered, when it hits one, two, all?

I guess this comes down to a question of who owns the message, the sender or the recipient? Home routing allows the recipient to delegate the job of message handling to their operator instead of their handset thus allowing the recipient to take advantage of all sorts of value-add services.

The business case for these systems that came up time and time again was SPAM filtering. The network wants to protect their subscribers from nefarious senders of unsolicited messages which generally originate from off the network.

I starting to believe that this is perhaps a bigger threat. The move by the Spanish operators to outlaw the use of alphanumeric originators, except those that have been pre agreed and do not originate off-net appears to be conducted under the banner of SPAM prevention.

The evidence I hear around the industry of legitimate messages being quietly discarded when sent to certain networks because of 'quality' issues is not the fault of home routing but more likely overly simplistic SPAM filters.

The subject of alphanumeric originators, network interconnects and the operators general distrust of A2P (application to person) traffic is a topic I will be returning to as I believe this is a more fundamental issue affecting us that could very quickly get out of hand.

Coming back to SMS home routing, I believe ultimately we have to accept it as an inevitability. As one operator contact told me, "I don't like it, but because xxx have done it we're going to have to as well". And let's be honest there are some really cool things it allows my network to do for me.

Much as we'd like not to, we have to accept that we faced with an oligopoly whose interests are not necessarily aligned with ours. The concept of delivery receipts and the knowledge they give us about the status of a communication is changing and we going to have to swallow that pill and adapt our services accordingly.

The rules have changed, but innovation thrives in times of adversity

Mobile working to Cannes

I decided to take the train to Global Messaging 2008 in Cannes,France. Last year I did it overnight which was sort of fun. This year I'm heading down during the day and seeing whether it's possible to do a days work while I spend the day travelling.

So armed with a my Sony Vaio, Vodafone USB Modem and BlackBerry I attempted to spend the day travelling but so that no one would notice.

3 trains required (if you ignore 2 stop on the RER in Paris):

1. Nottingham to London St Pancras
2. London St Pancras to Paris Gare de Lyon
3. Paris Gare de Lyon to Cannes

Nottingham to London - East Midlands Trains

This first leg is well known to me. I was able to use the power point to keep my laptop battery topped up but no WiFi and mobile coverage is shocking as usual. I've pretty much got used to not being able to do anything connected while on the train to/from London. That said my BlackBerry does an excellent job of exchanging emails when coverage allows.

Same goes for voice calls. Pointless trying to have a conversation.

It's always perplexed me as to why the mobile network operators have not invested in coverage for train lines. Little sardine cans of punters desperate to be communicate or consume content because there's nothing else to do.

conclusion: 3/10 great to have power, connectivity shocking

London to Paris - Eurostar

Aarrgh no power. Luckily I've invested in another battery for my laptop as the spare battery that came with my now 2 year old Vaio couldn't make it to Paris (the original battery gave up long ago). I guess though that a lot people's laptops will survive the 2hr journey so this shouldn't be too much of a hindrance.

I remember when Eurostar first opened and the trains seemed so glamorous, now they just seem tired and sadly lacking in basic facilities. Power points being one.

If they are trying to provide a business service then power is a must and I think WiFi would be a key asset. GNER (now National Express) seem to have done a great job on the East Coast Line, why not a premium route like London - Paris.

However this was when the Vodafone modem came to the fore, especially on the French side. I got 3G coverage almost all the way from Calais to Paris. I actually managed to use some web applications. Properly review some emails, respond completely. It was like being at my desk at home.

Voice calls no problem.

conclusion: 7/10 connectivity pretty good though WiFi would be beter, could really do with power

Paris to Cannes

This was going to be the real test. The first 2 legs are less than or about the 2 hour mark. If all else failed you could probably catch up with things at a coffee shop at one of the stations. 5 hours of whizzing through the French countryside was really going to put this experiment to the test, and it did.

I had high hopes for this train trip. The French being well known for their super fast train service, tearing round the country brimming with socialist pride. I imagined it would be well setup for the travelling business person.

I booked way ahead so travelled first class and was greeted by a nice big seat on the top deck (thought that might be better for mobile coverage) with my own power point.

No WiFi again, but as we left Paris I was hopeful as 3G coverage remained but this soon gave way to 2G, which in turn gave way to very patchy indeed. For large chunks of the journey it was unworkable. Especially as it seemed to stop connecting automatically, maybe that's a roaming setting.

Voice calls were also tricky, especially given they seem to be frowned upon in the carriage. Nothing like a mobile phone etiquette faux-pas in a foreign country to endear yourself with your fellow passengers.

conclusion: 6/10 power and seat great, data coverage varied

So is it workable? Certainly Eurostar, though more on the French side than the British. Given they've shortened that bit recently, I guess it's less of a problem ;-). The key at the moment is to know your route and know how to work round it's foibles.

Having used the WiFi on the East Coast (London to Edinburgh) line it's difficult to understand why other rail operators aren't putting it in across their fleet. You look in any rail carriage these days and the number of laptops as just mushroomed. Couple that with WiFi capable phones and PDAs and there's an army of people that would be quite prepared to pay an additional charge for WiFi while they travel.

I might be getting a little carried away, but I wonder if the government has any view on the national productivity hit of having people stuck on trains, unable to access their work. Could a government subsidy drive adoption as well as our GDP?