Agile Solutions Provider and the enterprise link South Africa still has to trust

Summary

• Agile Solutions Provider is best understood as a South African wholesale and business-connectivity provider, trading as AgileSP, whose public record points to IP transit, national long distance backhaul, dedicated internet access, colocation, remote peering and DDoS protection rather than a mass consumer broadband proposition.
• The economic unit to watch is the enterprise connectivity link: a contracted path between a customer site, data centre or network and AgileSP's own AS328748, with support, routing, upstream capacity and fault handling bundled into the price.
• The public evidence supports the thesis that local support, upstream diversity and accountability can matter more than headline speed, but it does not prove AgileSP's revenue, margins, customer count, private SLA performance or all customer outcomes.

The first sign that a connectivity supplier has become important is not a logo on a procurement deck. It is the phone call during an outage. A South African finance team cannot reach Microsoft 365. A logistics firm can see orders waiting in a cloud application but cannot reliably process them. A small ISP's customers blame the retail brand, while the actual strain sits deeper in transit, peering, backhaul and international cable capacity. At that moment, the useful link is not the cheapest line on a comparison table. It is the link attached to a reachable network team, a route plan that can move around trouble, and a contract that gives the buyer someone accountable when the public internet feels abstract and distant.

Agile Solutions Provider, better known on its own pages and network records as AgileSP, is a compact example of that South African enterprise-connectivity problem. Its public website describes the company as an African internet service provider based in South Africa, with a focus on carrier-grade IP transit and backhaul connectivity across Africa and Europe. Its PeeringDB record identifies Agile Solutions Provider (PTY) LTD as the organisation behind AS328748, gives AgileSP as an alias, points to the company website, and lists a technical footprint across South African and European exchanges and facilities. Its own PAIA manual gives the legal name Agile Solutions Provider (PTY) LTD, registration number 2019/118106/07, and says the company provides internet services and solutions. That is enough to place the business in the regional ISP and wholesale-connectivity market, but it is not enough to turn every marketing claim into a verified operating metric.

The judgement question is narrower and more useful. What does a South African enterprise, ISP, WISP, hosting provider or cloud-dependent business really buy when it buys an enterprise connectivity link from AgileSP? It buys bandwidth, but bandwidth is only the surface. It also buys routing policy, diverse upstream relationships, local and international exchange reach, access to data-centre locations, a technical support path, an initial service term, billing rules, suspension and termination rules, and a set of risk decisions about who carries the operational blame when things break.

That makes AgileSP a useful company to examine because its public footprint is unusually explicit for a small ISP member. The website lists IP transit, national long distance, fibre to the tower, colocation, remote IX peering, dedicated internet access and DDoS protection. The AS328748 site publishes a peering policy and BGP community list. PeeringDB lists operational ports at South African and European exchanges, including NAPAfrica Johannesburg, Cape Town and Durban, JINX, CINX, DINX, AMS-IX, LINX LON1, LONAP, NL-ix, Frys-IX, Speed-IX and MOZIX, plus data-centre presence in Johannesburg, Cape Town, Durban, London and Amsterdam. The BICS case study says AgileSP needed to scale global connectivity infrastructure, required high-quality IP transit between South Africa, Europe and Asia, and worked with BICS for direct, low-latency and resilient routes.

The record is not equally strong on every question. AgileSP does not publish audited financial statements in the public sources reviewed. It does not publish retail tariffs for a simple monthly comparison. It does not publish a public customer list that can be independently reconciled to revenue. The PAIA manual names Dean Pillay as current member and managing member, while the public pages reviewed do not disclose a parent company. The right conclusion is not that the business is opaque in a suspicious way. It is that the strongest evidence is operational and contractual, not financial. AgileSP's public case has to be assessed through the shape of the product, the network records that can be observed, the supplier and outage evidence around South African connectivity, and the way enterprise links are priced and supported.

Company identity and public footprint

Agile Solutions Provider presents itself under three related names: Agile Solutions Provider, Agile Solutions Provider (Pty) Ltd and AgileSP. The main website is https://www.agilesp.co.za. The separate AS328748 site at https://as328748.net/ is the network-facing record for peering policy, BGP communities and the looking glass. PeeringDB records the organisation as Agile Solutions Provider (PTY) LTD and the alias as AgileSP. The PAIA manual records Agile Solutions Provider (PTY) LTD with registration number 2019/118106/07, describes it as a proprietary limited internet service provider, and identifies Dean Pillay as current member and managing member. The current website footer uses the Midrand office address at Unit 7 Tybalt Place, First Floor, York House, Waterfall Office Park, Bekker Road, Midrand, while the PAIA manual lists 5 Kilimanjaro Crescent, Blue Hills, Midrand, and a Wierda Park postal address. The address difference is not unusual for growing small companies, but it is a reminder that public readers should distinguish current operating contact pages from older legal manuals.

AgileSP's own pages also make several regulatory and industry-membership claims. The about page says the company is ICASA licensed, a member of ISPA, supportive of MANRS, a member of WAPA and a B-BBEE Level 1 contributor. ISPA's public member list independently places Agile Solutions Provider among small members and links the member to https://www.agilesp.co.za. The PeeringDB record independently supports the AS328748 network identity and gives contacts for abuse, NOC, technical and sales functions. The public evidence reviewed did not provide an ICASA licence certificate page directly from the regulator, so the ICASA point should be treated as an official company claim unless a regulator register is separately matched.

The company sells to other networks and to businesses with demanding links, not mainly to households comparing fibre bundles. On its about page AgileSP says it enables ISPs, WISPs, content providers and cloud operators through peering presence at African and European internet exchanges and capacity across more than five diverse subsea cable systems. The IP transit page adds mobile network operators, enterprises and international carriers to the buyer list. The dedicated internet access page names enterprises, corporates, financial services, hosting and cloud providers, media and broadcast, law firms and professional services. The remote IX page targets ISPs, WISPs, content providers, international networks and hosting providers. The fibre-to-the-tower page targets mobile operators, tower companies, WISPs and rural broadband operators.

That customer mix matters. A household pays for the visible last mile and asks whether a speed test looks close to the advertised plan. A business buyer pays for something less visible: fewer handoffs, clearer fault ownership, direct peering or transit choices, symmetric bandwidth, route options, access to public IPv4 and IPv6, the ability to run BGP when needed, a supplier who can explain a cable cut, and a support team that knows the customer's network rather than reading from a generic script. AgileSP's evidence record is strongest where those enterprise and wholesale features are described.

What AgileSP actually sells

AgileSP's public product set has six main pillars.

First, IP transit is the centre of the company story. The IP transit page describes AgileSP as an African wholesale connectivity carrier with carrier-grade IP transit built on infrastructure it says it owns, engineers and operates. The page says the product connects Africa to the world, lists west-coast, east-coast and pan-African cable systems, and markets a footprint of 13 exchanges and 15 or more data centres. It states that service can scale from 1 Gbps to 100 Gbps and beyond, supports full BGP routing tables, BGP communities, dual-stack IPv4 and IPv6, RPKI and route origin validation, blackholing, MANRS-aligned routing and 95th percentile billing based on five-minute samples. It also says AgileSP operates AS328748, manages upstream connectivity through Tier 1 carriers in London and Amsterdam, and maintains peering relationships at 13 internet exchange points.

Second, national long distance is the domestic backbone layer. The NLD page defines national long distance circuits as dedicated, high-capacity links between major metropolitan areas, naming Johannesburg, Cape Town and Durban as examples. It says the service is purpose-built for ISPs, wireless operators, mobile networks and enterprises, and that it provides predictable symmetrical throughput with hard SLA commitments. It lists capacities from 1 Gbps to 100 Gbps and beyond, diverse routing for mission-critical links, and on-net data-centre access. It also describes standard routes including Johannesburg-Cape Town, Johannesburg-Durban, Cape Town-Durban and Johannesburg-Port Elizabeth, with cross-border circuits available by quote.

Third, dedicated internet access is the business-premises product closest to the assignment's economic unit. AgileSP's DIA page contrasts shared broadband with a private, uncontended connection and says DIA provides a private symmetrical internet circuit, no shared contention, guaranteed throughput, prioritised routing and defined fault-response commitments. It lists availability from 10 Mbps up to 100G depending on location and access infrastructure, with IPv4 and IPv6 dual stack and BGP or static routing handoff. It describes the product as suitable for cloud-first businesses, financial services, hosting and cloud providers, media and broadcast users, law firms and professional services. It also says on-net DIA can be ordered at key South African data centres with same-building delivery and no external fibre required.

Fourth, remote IX peering lets a customer reach exchange points without building its own physical presence at each exchange. The remote IX page says an ISP, content provider or network operator can participate at an exchange without being physically present there. The customer connects to AgileSP, and AgileSP provides the exchange port. It lists South African exchange options such as NAPAfrica Johannesburg, Cape Town and Durban, JINX, CINX, DINX and Speed-IX, plus European exchanges including AMS-IX, LINX LON1, LONAP, NL-ix and Frys-IX. The page says route-server and peer paths can reduce latency and transit costs for traffic exchanged directly. It also says the service starts at 1G port size.

Fifth, colocation lets customers place equipment near the network. AgileSP's colocation page says it has colocation across carrier-neutral facilities in Johannesburg, Cape Town, Durban, London and Amsterdam. It lists Teraco JB1, JB3, CT1, CT2 and DB1; Equinix JN1 and LD8; Africa Data Centres sites; OADC sites; Digital Parks Samrand; xneelo Samrand; and Nikhef Amsterdam. It markets half-rack, full-rack and cage options, redundant power, physical security, cooling, remote hands, cross-connects and scalable power. The specific facility list is important because enterprise-link economics change when a customer is already in the same data centre. Same-building cross-connects can shorten lead times and reduce external last-mile costs.

Sixth, DDoS protection is sold as an add-on to connectivity. The DDoS page says mitigation happens at network level, upstream of the customer, using scrubbing infrastructure, BGP blackholing and BGP Flowspec. It says always-on monitoring identifies attack signatures and that mitigation can begin automatically, with post-event reporting and 24/7 NOC support. The public BGP community page supports the blackhole part by listing blackhole community values for global blackholing and upstream-specific options for WIOCC, BICS, NTT and Level3. That does not prove the quality of mitigation under attack, but it does prove that AgileSP publishes traffic-engineering controls relevant to DDoS response.

The enterprise connectivity link as the economic unit

The cleanest economic unit is an enterprise connectivity link. It may be a DIA circuit from an office to AgileSP, an NLD circuit between data centres, a remote peering handoff, or an IP transit commit from a network operator. The buyer is not just buying a speed tier. The buyer is buying a bundle of capacity, route access, support, contract accountability and supplier risk.

The bundle is expensive for five reasons.

The first cost is access. A link has to reach the customer's site, tower, data centre rack or network handoff. If the customer is in a facility where AgileSP is already present, same-building delivery may reduce the cost and time. If the customer is off-net, someone has to provide last-mile fibre, civils, wayleave work, cross-connects or third-party access. AgileSP's own terms acknowledge that local access circuits and related equipment can create cancellation costs, and the NLD and DIA pages repeatedly distinguish on-net locations from location-dependent delivery.

The second cost is committed capacity. A wholesale link must be sized for the customer's traffic pattern, not just a marketing number. AgileSP's IP transit page lists capacity tiers from 100 Mbps in the enquiry form up to 50 Gbps and beyond, and product language from 1 Gbps to 100 Gbps and beyond. The DIA page lists 10 Mbps to 100G depending on access infrastructure. The NLD page lists 1 Gbps to 100 Gbps and beyond. These ranges show the company is selling a ladder of commitments rather than a single packaged retail bundle.

The third cost is upstream and peering diversity. A customer can buy cheap bandwidth from a single path and suffer badly when that path is congested or cut. AgileSP's public product case rests on access to multiple subsea systems, European anchor points, public exchange membership and Tier 1 upstreams. The BICS case study reinforces that AgileSP needed high-quality IP transit to connect customers in South Africa with Europe and Asia, and that low latency, redundant connectivity and 99.999% uptime were part of the requirement. This is where the article's thesis is strongest. South Africa's geography makes international traffic risk a commercial issue. If a link depends on one coastal direction, one upstream or one congested route, the low monthly price can be misleading.

The fourth cost is people. AgileSP repeatedly sells support and NOC access as part of the service. The IP transit page says 24/7 NOC monitoring. The DIA page says defined fault response commitments. The outage blog says customers want to be able to call and speak to someone rather than only log a ticket, and it says AgileSP provides hands-on troubleshooting support and dedicated internetwork and ICT expert engineers. That is marketing copy and company-authored commentary, so it should not be mistaken for independent customer satisfaction data. But it describes the labour component the company wants buyers to value: technical support close enough to understand routing, cable failures and customer growth plans.

The fifth cost is contract risk. AgileSP's standard terms define installation, start-up, rental, monthly and usage charges. They say rental and monthly fees may be billed monthly in advance, usage charges monthly in arrears, invoices due within seven days, charges exclusive of VAT, and initial terms generally at least one year if not otherwise stated. They also say early termination during an initial term can trigger a charge equal to 50% of remaining rental or monthly fees, excluding any part tied to local access circuits, plus cancellation costs or expenses linked to local access circuits, related services or equipment. This is not a consumer-friendly churn product. It is a business-services contract, and that structure tells the reader something about supplier economics.

The enterprise link is therefore a recurring commitment. It can be priced by monthly rental, usage, port size, committed data rate, installation, upstream capacity, local access, data-centre cross-connects, add-on DDoS protection and support intensity. Without AgileSP's private quote books, the actual rand price is not visible. But the public terms and product forms show the pricing logic clearly enough: the value proposition rises when the customer has high downtime cost, complex routing needs, cloud dependence, multiple sites, or a need for a supplier with network engineering depth.

Pricing proxies and what they reveal

Because AgileSP does not publish a tariff card for all services, the public evidence has to use proxies. There are at least five useful substitutes.

The first proxy is the IP transit capacity ladder and billing model. AgileSP's IP transit enquiry form asks for capacity options from 100 Mbps through 50 Gbps and above. The same page says 95th percentile billing is used, measuring traffic in five-minute samples and discarding the top five percent of peaks before billing on the remaining 95th percentile value. That tells us the product is built around committed and burstable bandwidth economics. A buyer with uneven traffic can burst, but the bill is still tied to sustained utilisation. For AgileSP, revenue scales with customer commits, port speeds and actual use; cost scales with upstream commits, exchange ports, router capacity and backhaul.

The second proxy is dedicated internet access capacity. The DIA page says service runs from 10 Mbps up to 100G, depending on location and access infrastructure. It also says the connection is private, symmetrical and uncontended, and can use BGP or static routing. That separates DIA from residential fibre substitutes. A 100 Mbps consumer line and a 100 Mbps DIA circuit can look similar in a headline but very different in economics. DIA has dedicated capacity, engineering, monitoring, service commitments and often custom access cost. The buyer pays for predictability, not merely maximum burst speed.

The third proxy is the standard terms. Installation charges, start-up charges, rental charges, monthly fees and usage charges are all defined. Monthly recurring charges begin on service activation, usage charges can be billed after the fact, and early termination can impose a substantial cost. This is evidence that AgileSP's pricing model depends on recovering upfront deployment and access costs over a term. It also shows switching costs for the customer. Replacing a supplier may mean new access, new BGP setup, new cross-connects, new IP addressing arrangements and a financial charge before the old term expires.

The fourth proxy is PeeringDB port and facility evidence. PeeringDB lists AgileSP's public peering ports, including 20G at NAPAfrica Johannesburg and 10G at several exchanges such as AMS-IX, CINX, DINX, JINX, LINX LON1, LONAP, NAPAfrica Cape Town, NAPAfrica Durban, NL-ix and Speed-IX, plus 1G at MOZIX. Peering ports and facility presence are not customer prices, but they are visible cost commitments. Exchange ports, routers, cross-connects and colocation cabinets are part of the fixed base that a wholesale carrier must monetise across customers.

The fifth proxy is BICS's case study. BICS says AgileSP needed to scale global connectivity infrastructure for a growing customer base and required high-quality IP transit connecting customers in South Africa with Europe and Asia. It says AgileSP needed to deliver 99.999% uptime and support high volumes of priority traffic. That source is supplier marketing, not an independent audit, but it is still useful because it identifies upstream dependence and the kind of resilience AgileSP has to buy before it can sell resilience to its own customers.

Together, these proxies support the thesis. AgileSP is not competing only on headline speed. Its public product is a set of costly commitments: multiple locations, exchange ports, upstream carriers, engineering support, route controls, dedicated access and contractual accountability. The unresolved question is whether customers consistently receive the operational value promised. Public pages contain testimonials and company-authored case claims, but the sources reviewed do not provide enough independent customer outcome data to measure fulfilment across the base.

Revenue logic, costs and supplier dependence

AgileSP's likely revenue logic is recurring connectivity revenue with quote-based setup, capacity and usage components. The company can sell a direct DIA link to an enterprise, a transit commit to an ISP, a transport circuit to a network operator, remote peering to a network that wants exchange access, colocation to a customer that needs rack space and cross-connects, and DDoS protection as a protective layer. A single customer can buy more than one product. For example, a WISP might buy fibre to the tower for backhaul, IP transit for global reach, remote peering for local traffic, and DDoS protection for customer prefixes. An enterprise might buy DIA, colocation and protection for cloud traffic.

The fixed cost base is visible even when the financial statements are not. AgileSP has to maintain routers and optical interfaces, lease or own equipment, pay data-centre space and power, buy cross-connects, pay exchange or transport-related costs, maintain NOC labour, manage regulatory compliance, buy upstream transit, maintain software and monitoring systems, and absorb sales and account-management costs. It also has to carry enough spare capacity to make an uptime and burstability promise credible. A provider that runs every port hot may look cheap until the first cable failure or traffic spike.

Variable costs change with traffic, access method and customer complexity. A same-building cross-connect in a data centre is different from off-net fibre to a customer premises. A static-route DIA customer is different from a customer with BGP, IPv6 requirements, DDoS communities and multi-site failover. A small enterprise that wants 50 Mbps of reliable access is different from a regional ISP buying multi-gigabit transit. Customer support is also variable. The costly customer is not always the largest. A smaller customer with unclear internal IT responsibilities and recurring fault escalations can consume more support time per rand of revenue than a larger network with experienced engineers.

Supplier dependence is central. AgileSP's own BGP community list names upstream-specific traffic-engineering and blackhole options for WIOCC, BICS, NTT and Level3. The BICS case study confirms at least one major upstream supplier relationship. The product pages talk about Tier 1 carriers in London and Amsterdam and cable-system diversity. That is a strength if the company has negotiated good routes and can move traffic when trouble appears. It is a risk if pricing, capacity or service quality from upstreams changes faster than AgileSP can pass cost through to customers.

Subsea-cable dependence is a special South African operating risk. MyBroadband reported that on 14 March 2024, South Africa suffered major internet problems after multiple undersea cable failures, naming WACS, ACE, MainOne and SAT3, with Microsoft later saying the combination of west-coast and Red Sea cable impacts reduced Africa capacity and affected public internet and cloud providers. AgileSP's own blog used that outage to argue for multiple cable systems, east and west coast redundancy and customer support. A supplier cannot make the sea floor safe. What it can sell is capacity on different systems, route control, customer communication and enough engineering work to avoid making every cut a full outage for every customer.

This is where headline speed becomes a weak procurement tool. A 1 Gbps line that depends on a fragile path can become a 1 Gbps promise on paper and a bad business day in practice. A lower headline-speed link with better route diversity and support may be more valuable for a customer whose applications, payments or customer service depend on continuity. The public evidence does not prove that AgileSP wins every comparison. It does show why the comparison has to include resilience and support.

Customer dependence and switching costs

Enterprise connectivity contracts create two-way dependence. AgileSP depends on recurring customers to cover fixed network costs. Customers depend on AgileSP once routes, addressing, firewall rules, BGP sessions, support paths and failover plans have been built around the link. The terms and product pages make those switching costs visible.

The standard terms say a customer order creates an individual contractual relationship for a service term, generally a minimum of one year if not otherwise stated. They define service activation and acceptance, monthly billing, usage billing, payment deadlines, suspension rights and termination rights. They also state that IP addresses assigned by AgileSP must be used only with the internet service and returned when the service ends. That can matter for customers that host systems, run whitelists, publish DNS records, operate VPNs or advertise routes. Renumbering or changing upstream relationships is operational work.

The physical switching cost can be larger than the paperwork. A customer moving from AgileSP to another provider may need new cross-connects, new last-mile fibre, new router ports, new BGP sessions, new firewall policies, new monitoring thresholds and new support escalation paths. If the customer is off-net, local access cancellation charges may also follow. If the customer bought DDoS protection tied to BGP communities, changing providers means rethinking mitigation and emergency runbooks. If the customer colocated equipment in a facility partly to be near AgileSP, the replacement supplier's presence in the same facility matters.

This dependence is not necessarily bad. It is the reason enterprise services exist. A buyer wants the supplier to know its network well enough to fix problems quickly. The risk is lock-in without performance. The buyer should demand clear service descriptions, current facility availability, a documented escalation path, route-control options, a realistic view of access lead times, and clarity on what happens during cable failures, DDoS incidents and upstream congestion. AgileSP's public pages provide some of this language, especially around SLA, NOC, route controls and diverse paths, but private customer orders would decide the enforceable detail.

Competitors and substitutes

AgileSP's competitive set depends on the buyer.

For IP transit and wholesale network services, the alternatives include larger South African and regional providers, international carriers, data-centre-connected networks, and a buyer's own direct peering build. PeeringDB itself shows the crowded nature of exchange ecosystems. A network that can justify its own ports at NAPAfrica, JINX, CINX, DINX, AMS-IX or LINX may choose direct peering instead of remote IX. A smaller network may prefer AgileSP because it can access multiple exchanges through one operational relationship.

For dedicated internet access, alternatives include business-fibre products from large ISPs, carrier Ethernet from telecom operators, fibre network operators with ISP overlays, fixed wireless, 5G, satellite and multi-link SD-WAN. Cheap fibre and 5G can be good substitutes for non-critical traffic. They are weaker substitutes when the customer needs symmetric bandwidth, public routing, predictable restoration, on-net data-centre delivery, DDoS controls or a supplier who can explain path behaviour during a regional cable incident.

For national long distance and backhaul, alternatives include direct circuits from major fibre owners, dark fibre, lit wavelength services, microwave backhaul and leased capacity from carriers with national networks. AgileSP's value depends on whether it can combine backhaul with transit, peering and support in a simpler relationship. The public product pages say it can, but individual routes and pricing would decide each tender.

For DDoS protection, substitutes include global scrubbing services, CDN-based application protection, cloud-provider controls, on-premise appliances and upstream blackholing from other carriers. AgileSP's advantage, if delivered, is integration with the connectivity service and BGP controls. Its weakness, compared with some global specialist mitigators, may be lack of public attack-volume data, public mitigation benchmarks or named scrubbing partners.

The competitive point is that AgileSP does not need to be the cheapest supplier to make sense. It needs to be credible for customers whose cost of downtime, support delay or route instability exceeds the saving from a cheaper link. Public evidence supports a product designed for that buyer. It does not yet prove market share.

Regulation, geography and operational risk

South African connectivity providers operate in a demanding environment. They face telecom regulation, data-protection and interception-law compliance, power constraints, fibre-break risk, vandalism and theft in parts of the physical network, exchange and data-centre concentration, foreign-exchange exposure for international capacity and equipment, and subsea cable events that can change routing economics overnight. AgileSP's acceptable-use policy references South African legislation including the Electronic Communications and Transactions Act, the Electronic Communications Act, the Films and Publications Act, and RICA. Its PAIA manual says records are maintained under laws including companies, copyright, employment, tax, VAT, labour, RICA, basic conditions of employment and access-to-information legislation.

Geography is both an opportunity and a risk. South Africa is a regional connectivity hub with major carrier-neutral data centres and exchange points. AgileSP's public footprint in Johannesburg, Cape Town and Durban lets it sell into the country's main commercial and interconnection centres. Its European presence in London and Amsterdam, according to product pages and PeeringDB, extends the route story beyond domestic access. But South Africa's international traffic still depends on subsea systems and upstream routes. The March 2024 cable outages show how quickly a regional issue becomes a business-continuity issue for banks, payment providers, cloud users and ordinary broadband customers.

Operationally, the most important risk is not one cable or one port. It is the difference between claimed diversity and working diversity. A supplier can list many cable systems, exchanges and facilities, but resilience depends on real traffic engineering, purchased capacity, spare headroom, physical route separation, failover testing and customer-specific design. The public record can show that AgileSP has the ingredients for diversity. It cannot show every private capacity commit or failover result.

Another risk is scale. BICS's case study says AgileSP wanted capacity that could grow with a rising customer base. Growth is good for a wholesale ISP because it spreads fixed costs. Growth is dangerous if support, backhaul and upstream purchases lag demand. AgileSP's public pages repeatedly say the company can scale, but the independent public record reviewed does not provide utilisation data or financial capacity. That is one of the facts that would change the judgement if it became available.

Network-resource evidence and its boundary

Public network records are useful here because they are less promotional than ordinary service copy. PeeringDB records AS328748 for Agile Solutions Provider, lists the alias AgileSP, gives the AFRINIC IRR set AFRINIC::AS-AGILESP, marks RIR status as ok, provides abuse, NOC, technical and sales contacts, and lists operational public peering entries and facilities. The AS328748 site publishes a peering policy for Agile Solutions Provider and a BGP community table. The looking glass lists routers in Amsterdam, Cape Town, Durban, Johannesburg and London.

Those records prove that AgileSP has a public internet routing identity, publishes peering and traffic-engineering information, and appears in exchange and facility databases as a network with multiple interconnection locations. They do not prove the company's revenue, the commercial terms of each customer, the live utilisation of each port, the quality of every support interaction, the amount of private capacity bought from upstreams, or whether a specific customer link will survive a specific outage. They are evidence of capability and footprint, not a guarantee of outcomes.

The technical evidence also helps separate AgileSP from generic reseller language. A simple reseller might publish a marketing page and little else. AgileSP publishes AS-specific records, BGP communities, peering policy and a looking glass. The community table includes blackholing, local preference, prepending and suppression controls. The peering policy says AgileSP announces only prefixes legitimately assigned to Agile or received from Agile customers, maintains route objects, maintains PeeringDB, supports public and private peering, follows MANRS principles, addresses routing mistakes, and requires strict filtering on customer BGP sessions. The policy also says public peering has no minimum traffic volume, ratio or contract requirement, while customer BGP sessions receive strict prefix, AS path and max-prefix filtering. These are operational details that a network buyer can test in technical conversations.

There are also caution flags. The peering policy says AgileSP generally does not peer with route servers, while some PeeringDB exchange rows show route-server peer indicators. That may reflect differences between display fields, policy wording or exchange-specific practice. It is not enough to draw a negative conclusion, but it is the kind of discrepancy a technically sophisticated buyer should ask about before relying on remote peering design.

Unofficial and commercial signals

AgileSP's public record includes several customer-facing signals that are useful but not conclusive.

The homepage includes five-star testimonials naming Matthew Winder, Gersh Fritz and Dr. Kamran Khan. It says one customer valued low latency and support, another says AgileSP helped scale IP transit over more than four years, and another says Bluedot Data customers remained unaffected by subsea cable breaks because of AgileSP's topology and redundancy. The 2024 outage blog includes named quotes from TouchNet, Bluedot Data and Open Link Communications. These statements fit the thesis: customers value uptime, support and scaling, not only speed. But the statements sit on AgileSP's own site. They are not a statistically representative review base, and they cannot prove average service quality.

The BICS case study is more independent than AgileSP's own page but still commercial. It comes from an upstream supplier describing its customer. It supports the fact that AgileSP sought scale, resilience and routes to Europe and Asia through BICS. It does not prove AgileSP's whole upstream mix, margins or customer satisfaction. It is best used as supplier-context evidence.

The MyBroadband March 2024 outage report is independent market context rather than a direct AgileSP review. It shows that South African businesses can be affected by cable failures and cloud-service reachability problems. AgileSP's own blog then uses that public incident to make a case for multiple cable systems and support. The combination is useful: MyBroadband supports the problem, while AgileSP's public pages show how the company claims to solve it. The gap is proof of customer-by-customer outcomes during the incident.

There is a broader unofficial signal in the structure of the market. South Africa has a large set of small and medium ISPs, WISPs, hosting companies, regional operators and enterprise buyers. ISPA's list places AgileSP among small members, and PeeringDB shows active interconnection. That does not prove revenue size, but it supports the idea that there is demand for suppliers between mass-market broadband brands and global carriers. AgileSP appears to occupy that middle layer.

What would change the judgement

Several public facts would materially improve or weaken the assessment.

Audited financials or credible turnover data would show whether AgileSP's network footprint is supported by a healthy revenue base. A supplier can have a good technical design and still be constrained by cash, supplier credit or scale. Public margins would show how much room the company has to buy spare upstream capacity and maintain support labour.

Regulator confirmation of licences would strengthen the legal record. AgileSP's own about page says it is ICASA licensed, and the company publishes legal policies, but a direct licence match from a regulator source would be stronger.

Independent customer data would be decisive. Public renewal records, procurement awards, business reviews with technical detail, or outage postmortems could show whether AgileSP's support promises hold across multiple customers. A few testimonials are useful colour; they are not enough to measure average service.

More detailed network measurements would help. Public latency traces, route changes during cable failures, RPKI coverage and looking-glass tests over time could show whether the advertised diversity is being used effectively. PeeringDB and BGP community records show the map. They do not show live performance under stress.

Finally, explicit tariff examples would sharpen the economics. A buyer can learn from capacity ladders, billing model and terms, but actual rand pricing would let readers compare AgileSP with business fibre, carrier Ethernet, transit and remote peering substitutes. Without price cards, the article can identify pricing logic but not rank AgileSP by cost.

Visible public evidence

https://www.agilesp.co.za/ - Official homepage. Supports the company's trading name, services, contact details, testimonials, self-description as a South African internet and telecommunications provider, and public service categories.

https://agilesp.co.za/about_us/ - Official about page. Supports the description of AgileSP as an African ISP focused on carrier-grade IP transit and backhaul across Africa and Europe, B-BBEE Level 1 claim, regulatory and industry-membership claims, and current Midrand contact details.

https://agilesp.co.za/agile-sp-promotion-of-access-to-information-act-section-51-manual/ - Official PAIA manual. Supports legal name, registration number 2019/118106/07, proprietary limited status, internet-service description, Dean Pillay as current member and managing member, and legacy contact particulars.

https://ispa.org.za/membership/list-of-members/ - ISPA member list. Supports Agile Solutions Provider's listing as an ISPA small member and links the member to the AgileSP website.

https://agilesp.co.za/our-services/ip-transit/ - Official IP transit page. Supports the core product, capacity ranges, cable-system diversity claims, exchange and data-centre footprint, 95th percentile billing, 24/7 NOC monitoring, AS328748 reference, dual-stack support and routing-security features.

https://agilesp.co.za/our-services/dedicated-internet-access/ - Official DIA page. Supports the enterprise-link product, uncontended private bandwidth, 10 Mbps to 100G range, BGP or static routing handoff, cloud-first use case, SLA language and same-building data-centre delivery.

https://agilesp.co.za/our-services/national-long-distance/ - Official NLD page. Supports the national backbone and inter-city circuit description, 1 Gbps to 100 Gbps range, standard city pairs, diverse routing and on-net data-centre delivery.

https://agilesp.co.za/our-services/remote-ix-peering/ - Official remote peering page. Supports the remote IX product, exchange list, one-provider access proposition, 1G starting port size and peering economics around reduced transit and latency.

https://agilesp.co.za/our-services/colocation/ - Official colocation page. Supports the listed data-centre locations, colocation features, cross-connect logic and facility-driven cost base.

https://agilesp.co.za/terms-and-conditions/ - Official standard terms. Supports billing categories, installation and start-up charges, monthly and usage charges, initial service term, early termination charge, suspension rights, liability limits, IP address return and customer obligations.

https://agilesp.co.za/acceptable-use-policy-2/ - Official acceptable-use policy. Supports the abuse, legal-compliance and network-integrity controls attached to service use.

https://www.peeringdb.com/asn/328748 - PeeringDB record. Supports the AS328748 identity, AgileSP alias, AFRINIC IRR set, contacts, peering policy link, RIR status, exchange ports, data-centre facilities and public network footprint.

https://as328748.net/ - AgileSP network site. Supports the public network-facing service list, AS328748 context and links to peering policy, communities, looking glass and PeeringDB.

https://as328748.net/peering/policy/list/ - Peering policy. Supports selective peering policy, route-object maintenance, MANRS and BCP38 language, public and private peering support, public peering requirements and customer filtering statements.

https://as328748.net/bgp/community/list/ - BGP communities. Supports blackholing, upstream-specific controls, local preference, prepending and route suppression controls.

https://lg.as328748.net/ - Looking glass. Supports public router locations in Amsterdam, Cape Town, Durban, Johannesburg and London.

https://www.bics.com/agile-sp-case-study/ - BICS case study. Supports the upstream-supplier context, AgileSP need for scalable global connectivity, routes from South Africa to Europe and Asia, resilience, 99.999% uptime requirement language, BICS relationship and Dean Pillay quote.

https://mybroadband.co.za/news/internet/528897-massive-internet-problems-in-south-africa.html - MyBroadband outage report. Supports the March 14, 2024 South African connectivity disruption, WACS, ACE, MainOne and SAT3 failures, cloud-service impact and public-internet dependency context.

https://agilesp.co.za/2024/04/02/the-agile-solution-to-internet-downtime/ - AgileSP outage commentary. Supports AgileSP's own explanation of subsea cable redundancy, customer quotes, data-centre presence claims, support emphasis and the company's framing of enterprise resilience during the March 2024 cable incident.

Bottom line

Agile Solutions Provider's public record supports a cautious positive thesis: for the right South African customer, the enterprise connectivity link is worth paying for when it includes local support, route engineering, upstream diversity and accountable service terms. AgileSP is not simply selling speed. It is selling a managed path through South Africa's interconnection geography and international-capacity risk.

The evidence is strongest on identity, service mix, technical footprint, contract structure and upstream-resilience logic. It is weaker on prices, revenue, margins and independently measured customer outcomes. That means the commercial judgement should not be "AgileSP is always better than cheaper fibre." The better judgement is that AgileSP belongs on shortlists where downtime has a measurable business cost, where the buyer needs BGP or data-centre interconnection, where local support labour matters, and where a renewal discussion has to ask who will answer when the next cable incident turns a headline-speed promise into a real operational test.