Smppcube vs enterprise wholesale platforms (Alaris / HORISEN class)

This is the comparison where we have to be most careful, because the honest answer contains a sentence our competitors would never write: for a real wholesale carrier, the enterprise platform is worth every dollar and we are not the right purchase. Alaris, HORISEN and the rest of that class are serious telecom software, built by people who have spent decades inside the settlement disputes and fraud patterns of the international messaging market, and they are priced accordingly, somewhere between 20,000 and 150,000 USD before anyone has said the word “maintenance”. The problem is not that they are expensive. The problem is that most of the people quoting them are not doing the job the expensive part was built for. Below: what that class is genuinely designed to do, the ninety percent that overlaps with a 6,400 USD licence almost line for line, the ten percent that does not and what it is really worth, the price shapes and what they do to your margin, and a straight answer on which one your stage of business actually needs.

What the enterprise platforms are actually built for

Look at what these products optimise and the intended buyer is unmistakable. They are built for trading traffic, not for sending it.

A wholesale carrier does not have five routes, it has three hundred. It buys the same destination from eleven vendors at eleven prices with eleven qualities, sells it to forty customers at forty prices, and lives on the difference. The margin on any single message is a fraction of a cent, so the entire business is a question of whether your software can pick the right vendor for this message, this second, and then prove it six weeks later when the vendor’s invoice disagrees with your records by 41,000 USD.

That is the job. Everything expensive in the enterprise platform follows from it: least-cost routing that re-ranks itself as prices and quality move, a settlement engine because you owe money to the people who also owe money to you, a dispute workflow because the reconciliation will fail, fraud detection because a fraction of your inbound traffic is actively trying to steal from you, signalling because you interconnect at the carrier layer, and a NOC because a bind dropping at 03:00 is not an inconvenience, it is a contractual penalty.

Now the useful question. Read that paragraph again, and count how many sentences describe your week. If the answer is “most of them”, stop reading, go and buy the enterprise platform, and negotiate hard on maintenance. If the answer is “none of them”, the rest of this article is about the 30,000 to 100,000 USD of engineering you are about to buy for a job you do not have.

The ninety percent that overlaps

Here is the uncomfortable part for a vendor at their price point. Strip out the wholesale trading apparatus, and what is left is a messaging platform, and messaging platforms have converged. The list below is not a claim that we are cleverer than they are. It is a claim that this layer is well understood, and that a 6,400 USD licence and a 90,000 USD licence do it in ways your clients cannot tell apart.

Multi-tenancy with a reseller tree. Clients under resellers under you, with isolation that actually holds: a reseller sees their own tree and nothing above it. Both do this. It is table stakes at either price.

Rate cards per client, per route, per destination. Prefix-level pricing, effective dates, currency per card, bulk import. Both. This is the feature people assume is the expensive one, and it is not.

Prepaid, postpaid and credit control. A balance that is enforced on the hot path, a credit limit that is honoured, a refund on a hard failure, and idempotency so a retry never charges twice. Both, and the correctness question here is engineering discipline, not price.

Invoicing. Recurring, prepaid, postpaid, multi-currency, tax lines, PDF out. Both.

SMPP server and HTTP API in, SMPP and HTTP out. Your clients bind to you, you bind to your carriers, delivery receipts correlate end to end. Both.

White-label portals. Per-tenant logo, colours, domain, sender-visible branding. Both.

Reporting your client can read and your accountant can reconcile. Traffic, delivery, spend, margin by client and route, exports. Both.

Routing rules. Per-destination, per-sender, failover, throughput caps, blackout windows. Both, with a caveat this article will return to: “routing rules” and “least-cost routing across three hundred vendors” are different products wearing the same word.

Omnichannel. SMS, and increasingly WhatsApp, RCS and voice. Both, and here we would gently note that a platform built in 2003 for wholesale SMS interconnect tends to treat WhatsApp as a module purchased separately rather than as something in the same platform.

Nine rows. On any of them, a demo of the two products side by side produces a shrug. That is the ninety percent, and it is what your reseller clients log into every day.

The ten percent, honestly

Now the other side, because this is where the money is and where we lose the deal on merit. Seven capabilities. For each one, we have written the test that tells you whether you need it, and we would ask you to answer the test rather than the feature name, because feature names are designed to make you want them.

Least-cost routing at trading scale. Not “send to vendor B if vendor A fails”, which is a failover rule that any platform has. This is a live ranking across hundreds of vendor routes per destination, weighted by price, delivery rate, latency, and sender-ID survival, re-scored continuously, with margin floors so it never routes you into a loss. The test: do you buy the same destination from more than five vendors and change the mix during the day? If you have three carrier contracts and a preferred route, you do not need this, you need failover, and you have it.

Interconnect settlement and reconciliation. Your CDRs against their invoice, matched, with the deltas surfaced, aged and pushed into a dispute workflow with a paper trail that survives an audit. This is genuinely hard software and it is genuinely valuable, because at wholesale volumes a one percent reconciliation gap is a salary. The test: do you receive invoices from vendors that you must check, and does anyone dispute yours? A retail A2P reseller settles with nobody. They buy on contract and sell prepaid. There is nothing to reconcile.

Wholesale fraud detection. Artificially inflated traffic, SIM boxes, grey routes, revenue-share fraud on the A2P termination side, faked delivery receipts from a vendor who is dumping your traffic. This class of platform has pattern libraries built from twenty years of watching it, and that library is not something you can write in a sprint. The test: does traffic arrive at your platform from parties whose incentives you do not control? If your inbound is your own clients on contracts you signed, your exposure is a different and much smaller problem.

Carrier signalling. SS7 and SIGTRAN, HLR lookup at wholesale volume, deep number-portability integration. Real, and out of scope for us: Smppcube does MNP at the level a reseller needs, not at the level of an operator holding an SS7 point code. The test: do you have, or are you buying, a signalling interconnect? If you have never priced an SS7 link, this row is not for you.

A route-testing probe network. Real SIMs in real handsets in real countries, sending test traffic through your vendors on a schedule so you learn a route has degraded before your customer’s complaint does. This is a physical asset and an operational commitment, and it is one of the more defensible things in the enterprise offer. The test: do you resell routes whose quality you cannot otherwise verify? Note that you can also buy probe testing as a standalone service from several vendors and point it at any platform, including ours, which is worth knowing before you buy a whole platform to get it.

A 24/7 NOC with a contractual SLA. Humans, awake, with penalties attached to their answer time. This is not software at all, and it may be the single most valuable line in the quote. It is also the line to interrogate hardest: ask what the SLA pays out, ask for last year’s actual availability, and ask whether the NOC covers your configuration or only their platform’s health. The test: what does one hour of downtime cost you, and who is currently awake for it?

Multi-datacentre active/active at carrier volume. Proven at tens of thousands of messages per second, with a support contract that stands behind the tuning. Smppcube runs active/active, and we are straightforward about the difference: ours is proven at aggregator volumes, not at tier-one carrier peak, and the number that matters is not the architecture diagram, it is who is contractually responsible when it does not hold. The test: is your peak-second throughput a number anyone has ever measured, or a number in a proposal?

Read the seven tests together and a pattern falls out. Six of them are the same question: is somebody else’s money and somebody else’s traffic flowing through your platform in a relationship you have to police? That is the wholesale business, and the ten percent is the cost of policing it. If you buy on contract and sell retail, you are being quoted for a police force you will never deploy.

Price shapes, and what each one does to you

Nobody publishes these numbers, so treat the ranges below as what operators report rather than as a price list, and make the vendor put their own numbers in writing. The shapes matter more than the amounts.

Shape	Typical range	What it does to you
Perpetual licence, per module and capacity	20,000 to 150,000 USD	Big up front, then annual maintenance forever. Capacity tiers reprice you upward as you grow.
Annual maintenance	15 to 22 percent of licence, per year	3,000 to 33,000 USD a year. Usually not optional if you want updates or support.
Subscription	2,000 to 10,000 USD a month	72,000 to 360,000 USD across three years, owned at the end: nothing.
Revenue share, per message	0.00002 to 0.0001 USD, monthly minimum common	Scales exactly with your success. At a 0.0030 USD spread that is 0.7 to 3.3 percent of gross margin.
Implementation and professional services	10,000 to 50,000 USD	Not optional. These platforms are configured by their people, not installed by yours.
Change requests after go-live	Day rate	Every workflow that is not theirs is a quote.

Three observations we would want a friend to hear.

First, the licence is rarely the biggest number. Maintenance at 18 percent of a 90,000 USD licence is 16,200 USD a year, which crosses the entire Smppcube licence in five months and never stops. Ask for the three-year total including maintenance and services before you compare anything.

Second, capacity tiers and per-module pricing are where the quote grows after you sign. The second datacentre, the second protocol, the WhatsApp module, the next throughput band. Each one is defensible on its own and each one arrives on the day something good happens to you, which is the same mechanic we described in the rented-panel article wearing a much better suit.

Third, the revenue-share shape is the friendliest and the most expensive. It feels like alignment. It is a percentage of your gross margin, forever, computed on your revenue rather than your profit, and it rises every time you sharpen a quote to win a client. It looks tiny per message because it is designed to.

The comparison, side by side

	Enterprise wholesale platform	Smppcube v9
Built for	Wholesale traffic trading	Retail A2P, resellers, aggregators
Up-front cost	20,000 to 150,000 USD, or 2,000 to 10,000 USD a month	6,400 USD, once
Annual maintenance	15 to 22 percent of licence, forever	None
Implementation	10,000 to 50,000 USD, their team	Guided install, days
Multi-tenant reseller tree	Yes	Yes
Rate cards per client, route, destination	Yes	Yes
Prepaid, postpaid, credit control	Yes	Credit, wallet and auto routes
Invoicing, multi-currency	Yes	Prepaid, postpaid, recurring, multi-currency
SMPP server and HTTP API	Yes	Yes
White-label portals	Yes	Yes, per tenant, own logo and domain
Failover and routing rules	Yes	Yes
Least-cost routing across hundreds of vendors	Yes, quality-weighted, continuous	No, and we will not pretend otherwise
Interconnect settlement and dispute workflow	Yes	No
Wholesale fraud, AIT and SIM-farm detection	Yes, mature pattern libraries	No
SS7 and SIGTRAN signalling	Yes	No
Route-testing probe network	Yes, or bundled	Buy separately if you need it
24/7 NOC with contractual SLA	Yes, priced in	Support contract, no carrier SLA
Active/active	Yes, carrier scale	Yes, aggregator scale
WhatsApp, RCS, voice	Usually separate modules	In the same platform
AI features	Cloud, if any	Offline-capable, admin-selectable
Source code	Almost never	Delivered
Customisation	A change request at a day rate	You own the code
Time to a working client portal	Weeks to months, their calendar	Days, yours
Asset at the end of year three	A licence and a maintenance bill	The platform

The grid is only useful if you read it in two halves. Everything down to the failover row is a tie, and it is the half your clients touch. Everything from least-cost routing to the probe network is a clean win for them, and it is the half a retail reseller never opens. The bottom four rows are ours, and they are the ones that are hard to feel in a demo and impossible to ignore in year three.

Three years of cash, for a business that is not a wholesale carrier

Assume a regional aggregator: 3,000,000 messages a month, ten to forty clients, three carrier contracts, no interconnect settlement, no SS7. Middle of the ranges above, one datacentre, no change requests, which is generous.

Line item, 3 years	Enterprise platform (licence shape)	Enterprise platform (subscription shape)	Smppcube v9
Licence	60,000 USD	None	6,400 USD, once
Maintenance at 18 percent per year	32,400 USD	Included	0 USD
Subscription, 4,000 USD per month	None	144,000 USD	0 USD
Implementation and services	25,000 USD	15,000 USD	Included
Servers, 60 to 120 USD per month	2,160 to 4,320 USD	2,160 to 4,320 USD	2,160 to 4,320 USD
Three-year cash	119,560 to 121,720 USD	161,160 to 163,320 USD	8,560 to 10,720 USD
Owned at the end	A licence, plus year four’s maintenance	Nothing	The platform, perpetually
Cost of the fourth year	10,800 USD maintenance, plus servers	48,000 USD, plus servers	The servers

The ratio is roughly twelve to one, and we would rather you distrusted it than took it on faith, so here is what it is not. It is not a claim that the enterprise platform is twelve times worse. It is a claim that for this business, the extra 110,000 USD buys the seven capabilities in the ten percent section, and this business uses none of them. Change the assumption to a wholesale carrier settling with sixty vendors, and the same 110,000 USD stops being overhead and starts being the reconciliation engine that recovers more than that in disputed CDRs. The number does not change. The job does.

Which one your stage of business needs

Three shapes of business, and the honest answer is different for each.

You resell A2P to end customers. Two to five routes, contracts you signed, clients who buy prepaid credit. Your margin comes from selling messages. You need the ninety percent, in full, run well, on your own server, and you need exactly none of the ten percent. Buy the licence, spend the difference on finding client number thirty, and revisit this page if your traffic mix ever turns wholesale.

You are a regional aggregator. Ten to a hundred clients, a few million messages a month, three to eight carrier contracts, some resellers under you, and the occasional route swap when quality slips. This is the ambiguous middle, and it is where most of the bad purchases in this market get made, because the enterprise sales motion is very good and the word “carrier-grade” is very persuasive at 3,000,000 messages a month. The test that settles it: do you trade routes, or do you sell messages? Swapping a vendor twice a quarter is not trading, it is procurement, and procurement does not need a least-cost routing engine. Almost everyone in this bracket is our customer, and the ones who are not know exactly why.

You are a wholesale carrier or an interconnect trader. Hundreds of vendor routes, settlement both ways, disputes, fraud arriving from parties you did not choose, signalling, and a peak second that has a real number attached to it. Buy the enterprise platform. Negotiate the maintenance percentage, cap the day rate, and get the SLA penalties in writing. We are not your platform and it would be a disservice to suggest otherwise.

When we would tell you to buy theirs

Four cases, plainly.

You settle with vendors. The moment money flows in both directions and an invoice can be wrong, you need a reconciliation engine with a dispute trail, and building one is not a project you want. This single line justifies the entire price for the businesses it applies to.

Fraud is a P&L line, not a worry. If AIT, SIM farms or grey routes have already cost you money this year, buy twenty years of pattern library. It is cheaper than the education.

Somebody’s contract requires it. A tier-one interconnect that demands SS7, a regulator that demands a specific reporting shape, a customer whose procurement requires a named vendor with a NOC and penalties. These are not technical arguments and you will not win them with a better architecture.

You have no operations capability and will not build one. Owning the source code is worth nothing if nobody on your side ever reads it. The enterprise deal bundles a NOC, an implementation team and someone to shout at, and if that is genuinely your gap, you are buying people and the software is the wrapper. Be clear that is what you are buying, and price the people accordingly.

The middle path nobody quotes you

There is a third option the enterprise sales cycle will not put on the table, and it is what we would do.

Buy the ninety percent now, at 6,400 USD, and buy the ten percent later, if the business asks for it. Both worlds speak SMPP, so this is not a one-way door. The migration, when and if it comes, is a re-bind and a data export, not a rewrite, and it is far cheaper than the reverse mistake of buying a wholesale platform for a retail business and discovering in year two that you paid 120,000 USD for multi-tenancy and invoices.

Two rules make the future move cheap, and both cost nothing today. Own your domain and your API hostname from day one, so your clients are integrated against you rather than against whoever is currently behind the record. And keep your CDRs exportable in a shape another system can eat, which you can, because the database is yours and the source code is on your disk.

There is also a coexistence pattern worth knowing, because several operators run it. The enterprise platform, or a specialist service, handles the wholesale layer: settlement, probes, fraud. Smppcube runs the retail layer: your resellers, their clients, the portals, the rate cards, the invoices, the WhatsApp and RCS and voice your clients keep asking for. They bind to each other over SMPP like any other route. You are not obliged to buy one platform to do two jobs, and the class of platform in this comparison has historically been much better at the first job than the second.

Deciding this in an afternoon

Four questions. Write the answers down before the next demo, because a good demo is designed to change them.

First, do you trade routes or sell messages? If your margin comes from picking the cheapest acceptable vendor per message across a large vendor set, you are trading, and you need the engine. If your margin is a spread on a route you contracted last quarter, you are selling, and the engine is furniture.

Second, does anyone send you an invoice you have to check? If yes, you need settlement, and settlement alone can justify the quote. If your only invoices are the ones you send, there is nothing to reconcile and you are being priced for reconciliation.

Third, what is the three-year total including maintenance and services? Not the licence. Add 15 to 22 percent a year, add implementation, add the modules that are not in the base, add the capacity tier you will hit in month twenty. Then put 6,400 USD next to it and look at the gap. If the gap is smaller than the value of the seven capabilities to your actual business, buy theirs with our blessing. If the gap is 100,000 USD and you cannot name two of the seven you would use, you have your answer.

Fourth, when they say no, what do you do? Every platform eventually meets a workflow it does not have. With a licence and a day rate, the answer is a quote and a queue. With source code on your own server, the answer is a developer and a Tuesday. That difference is invisible in the demo and it is the whole story of year three.

Then decide, and be at peace with it. If you are a wholesale carrier, go and buy the best enterprise platform you can negotiate, and read this article again when someone tries to sell you a 6,400 USD licence for a settlement problem. If you are not, then what you are being quoted is a superb answer to a question you have not been asked: a police force for money that is not moving in both directions, a routing engine for vendors you do not have, and a NOC for a peak second nobody has measured. The ninety percent that runs your business is already written, already multi-tenant, already billing, and it costs 6,400 USD once, with the source code in your hands and a fourth year that costs the price of a server.