Speed Star Dot Net and the Microeconomics of Bangladesh's Broadband Edge

Speed Star Dot Net and the Microeconomics of Bangladesh's Broadband Edge

Speed Star Dot Net is not important because it is large. It is important because it is small, visible, and positioned at the point where the broadband economy in Bangladesh becomes concrete. It owns or controls a registered autonomous system, advertises a small block of portable addresses, sells low-cost fixed broadband in a local Dhaka market, and appears to publicly depend on a concentrated upstream path. From this combination, the company illustrates a broader economic fact about small network operators in Bangladesh: the retail ISP may hold the customer relationship, but much of the economic power lies above, in the stewardship of addressing resources, domestic backhaul, access to international gateways, accessibility to content caches, license compliance, and route visibility.

Public registries identify the network as Speed Star Dot Net, operating AS149806. The APNIC registry links the organization to ORG-SSDN2-AP and the address KA-119/A, Kuril, Dhaka 1229, Bangladesh. Public BGP views show a small Bangladeshi access network with 512 IPv4 addresses originating from 103.186.216.0/23 and associated /24s, as well as IPv6 space under 2400:7520::/32, with RPKI-origin valid status in observed sources. Its website, under the Speed Star.Net brand, offers local broadband plans from 500 Tk per month for 10 Mbps up to 3,500 Tk per month for 180 Mbps, promoting BDIX connectivity, Facebook and YouTube caching, fiber optics, shared and dedicated bandwidth, dedicated IP options, and 24/7 support.

The economic interpretation is finer than a company profile. Speed Star Dot Net is an edge operator whose public assets are scarce but significant: an ASN, portable IP resources, a license footprint, a local retail brand, and a route entity visible to global collectors. These assets do not make the company autonomous in a strong sense. They give it an option on autonomy. The public routing table still shows a concentrated dependency surface, with Alfaz Network visible as the upstream path in major BGP views. Alfaz itself depends on larger providers, including Summit Communications, Level3 Carrier and Rego Communications in the BGP.tools view. Speed Star thus fits into a nested bargaining chain: local customers pay Speed Star; Speed Star depends on an upstream aggregator; the aggregator depends on larger national and international mid-network providers; and the regulatory system governs which classes of operators can sell access, provide national transmission, and supply international connectivity.

The identity problem: one operator, multiple names

The canonical registry identity is Speed Star Dot Net. APNIC records list ORG-SSDN2-AP as "Speed Star Dot Net", a LIR in Bangladesh, with an address in Kuril, a telephone number and an administrative email. The APNIC inet6num record for 2400:7520::/32 gives the netname SPEEDSTARNET-BD, describes the holder as Speed Star Dot Net, marks the allocation as portable, and associates the lower and route maintainer with MAINT-SPEEDSTARNET-BD. Hurricane Electric's rendering of the APNIC aut-num entity identifies AS149806 as SPEEDSTARNET-AS-AP, with the description Speed Star Dot Net and the country Bangladesh.

The operational identity is less clear. The public website presents the company under the Speed Star.Net brand. The ISP Association of Bangladesh registers use "Speed Star. Net", list the websitewww.speedstarbd.netand indicate an Upazila/Thana ISP license number, a trade license number, a BIN, a TIN, a mobile number and an email address. The BTRC's Upazila/Thana ISP list also registers "Speed Star.Net" at Ka-119/1, Kuril, Vatara, Dhaka-1229, with the same license number as the ISPAB member profile.

This naming ambiguity is typical of small access providers in fragmented broadband markets. The economically relevant question is not whether every punctuation mark matches. It is whether the records point to the same operational asset. Here they do. The APNIC organization, the BGP identity, the ISPAB member register, the BTRC license list, and the website all converge on the name Speed Star, the Kuril/Vatara geography, the speedstarbd.net domain, and the same local ISP activity. Minor variations such as "Dot Net", ".Net" and ". Net" should be treated as spelling and registry format noise, unless legal filings prove distinct corporate entities.

There is no public evidence in the examined sources that Speed Star Dot Net is a subsidiary of Coronet Corporation Limited, or that Coronet is a successor owner. Directory clues provided by users mentioning CORONET CORP LTD, Coronet Corporation Limited and Asia Pacific Network Information Centre are economically significant, but they should not be over-interpreted. Coronet is a major IIG and IP transit operator in Bangladesh with AS149765, substantial peering, and an official service portfolio aimed at service providers. Speed Star is a small AS149806 access network. The prudent interpretation is that both appear in the same Bangladesh routing and registry ecosystem. A direct business relationship is possible; a capital control relationship is not established by the public evidence reviewed here.

The active website adds another identity clue. It gives the company address as KA-119/A, Kuril, Vatara, Dhaka-1229, lists phone numbers +880 1724 808 454 and +880 1977 774 424, and usesinfo@speedstarbd.net. The footer states ©2022 Speed Star.Net and says the site is powered by Symbiosis ICT Solutions. This last line is a web production signal, not proof of ownership. It does not establish Symbiosis as a parent company, financier, or network operator.

The ownership register remains thin. The ISPAB member profile indicates that no director information is available, and its entry in the member list shows that no representative is registered. This does not mean the company lacks a controller; it means the association's public page does not disclose one. For infrastructure economics, this matters because beneficial ownership changes the bargaining model. An independent owner with a few thousand lines has a different cost of capital and risk profile. A local retail brand captive to a larger upstream operator has another. A reseller operating under an older brand has yet another. The registry trail proves network stewardship and operational identity; it does not prove ultimate control.

The service surface: cheap bandwidth, local trust, and built perception

Speed Star's website sells what small urban access ISPs in Bangladesh commonly sell: fixed residential and small-business broadband, packaged in speed tiers, with local content acceleration and support promises used to differentiate an otherwise commoditized connection. The headline claims are "trusted Internet service provider", 99% uptime, and 24/7 support. The advertised technical add-ons are BDIX connectivity, Facebook and YouTube caching, dedicated IP, full duplex, and fiber optics. The marketing language emphasizes local performance and local support rather than national scale.

The pricing grid is instructive. The advertised monthly plans are 10 Mbps at 500 Tk, 30 Mbps at 600 Tk, 40 Mbps at 700 Tk, 50 Mbps at 800 Tk, 60 Mbps at 1,000 Tk, 70 Mbps at 1,200 Tk, 80 Mbps at 1,500 Tk, 90 Mbps at 1,800 Tk, 100 Mbps with IP at 2,000 Tk, 130 Mbps with IP at 2,500 Tk, 150 Mbps with IP at 3,000 Tk, and 180 Mbps with IP at 3,500 Tk. Higher plans include a real or dedicated IP offering, while all advertised tiers include BDIX connectivity, FTP server access, and 24/7 uptime.

This pricing reveals several mechanisms. First, the advertised speed is not priced linearly. The 10 Mbps plan costs 50 Tk per advertised Mbps, while the 30 Mbps plan costs 20 Tk per advertised Mbps and the 180 Mbps plan is under 20 Tk per advertised Mbps. This does not imply that the operator buys wholesale bandwidth at these rates. Retail residential broadband is oversubscribed. The economic product is not guaranteed international capacity; it is a probability distribution of usable throughput over time. The operator sells enough peak-hour experience to avoid churn, while relying on statistical multiplexing, local caches, BDIX traffic, and the fact that not all subscribers saturate their line simultaneously.

Second, the "real IP" feature appears only on higher tiers in the public pricing grid. This is a rational response to IPv4 scarcity. Public BGP views show that Speed Star originates 512 IPv4 addresses. If the network has even a few thousand active users, public IPv4 addresses cannot serve as unique per-customer identifiers. The likely operational model is carrier-grade NAT or other address sharing for ordinary users, with public IPv4 monetized as a premium add-on for gamers, teleworkers, CCTV users, small offices, and customers running reachable services. The evidence does not reveal Speed Star's NAT architecture, but the combination of 512 IPv4 addresses, consumer ISP classification, and the real IP offering on higher tiers strongly indicates IPv4 scarcity as a pricing design constraint.

Third, BDIX and caching are not minor technical claims. They are economic instruments. A customer buying a low-cost local broadband line perceives value mainly through video, social platforms, gaming, messaging, and national content. If YouTube, Facebook, and local BDIX routes work well, the customer may perceive the connection as fast even when international transit capacity is constrained. The operator can therefore defend retail satisfaction without buying enough expensive upstream capacity to support the advertised throughput to all international destinations at all times. In low-ARPU, high-local-competition markets, cache hit rates and national exchange performance are margin levers.

Fourth, the language of public testimonials is local. A testimonial on the website says that the provider is reliable "around Kuril". That is a small phrase with economic significance. It suggests that the competitive battleground is not Bangladesh, nor even Dhaka, but a service territory at the neighborhood or building complex level. In such a market, brand trust comes from field repair times, payment flexibility, local technicians, building access, and the fact that the provider has already pulled fiber to an apartment block.

Geography and layer: an access ISP, not a national operator

The operational footprint visible in public registries is concentrated around Kuril/Vatara in Dhaka. APNIC gives the organization address as KA-119/A, Kuril, Dhaka 1229. ISPAB gives Ka-119/1, Kuril, Vatara, Dhaka-1229. The BTRC Upazila/Thana ISP list registers the same neighborhood and license number. The website's contact section uses the same family of addresses.

The license category matters. The ISPAB profile describes the BTRC license type as Upazila/Thana. The public BTRC list is also an Upazila/Thana ISP list. In Bangladesh, this is not the same economic position as an international gateway, a national ISP, a national transmission provider, or a tower/fiber infrastructure company. It is a localized access market authorization. It indicates last-mile retail service, not ownership of a national mid-mile network.

At the network level, BGP evidence supports the same conclusion. BGP.tools classifies AS149806 as an "Eyeball" network and lists one upstream, two peers, three originated IPv4 entries, and five originated IPv6 entries. Hurricane Electric shows a total of eight originated prefixes, all RPKI-valid in its view, and one observed BGP peer. IPinfo describes AS149806 as an ISP, classifies it as a consumer ISP, and notes no hosted domains in its dataset.

This is the profile of a local broadband edge. It originates customer-facing address space; it does not appear as a major transit provider; it does not host a visible ecosystem of domains; and it does not advertise a large inventory of routes. This does not mean the network is trivial. The last mile is the layer where retail money is collected. But the technical footprint indicates that Speed Star's value lies more likely in customer relationships, local distribution, and address stewardship than in wholesale network scale.

The visible network: AS149806 as a small stub system

AS149806 was allocated in the APNIC region in April 2022 according to public views from BGP.tools and IPinfo. Current public routing evidence shows a narrow address pool: 103.186.216.0/23, its constituent /24s, and five IPv6 /48 announcements under 2400:7520::/32. The APNIC inet6num entity registers 2400:7520::/32 as a portable allocated IPv6 block for Speed Star Dot Net. Both BGP.tools and Hurricane Electric show RPKI validity for the visible originated routes.

The term "stub AS" is economically useful. A stub network primarily originates its own customer traffic and does not provide wide transit to others. IPinfo tags include "stub AS" and describe the network as having a pronounced day/night rhythm, consistent with a consumer or eyeball network. BGP.tools calls it an eyeball network. These observations align with the residential and small-business broadband plans on the website.

A pool of 512 IPv4 addresses is small but valuable. In a mature market where IPv4 is scarce, 512 public IPv4 addresses can support customer NAT pools, public IP add‑ons, infrastructure, management, and business customers. The direct book value of the addresses may be modest compared to a large operator, but their option value is high for a local ISP: they enable independent routing, make upstream changes less disruptive, and allow the company to sell a differentiated "real IP" product. The IPv6 allocation is huge in address count, but its immediate monetization is weaker because many customer devices, applications, and support processes still depend on IPv4 reachability. IPv6 helps future‑proof the network and improve routing hygiene; IPv4 continues to power scarcity pricing.

The valid RPKI signal is worth noting. Many small networks fail on route entity and RPKI hygiene because the operational return is indirect. Speed Star's visible RPKI-valid status suggests that at least one competent administrator or an upstream support flow has aligned prefix origins with registry authorization. This reduces the probability of accidental route rejection by networks applying RPKI origin validation. It does not prove high-quality service, but it is a positive indicator of registry discipline.

The upstream signal is more binding. BGP.tools lists Alfaz Network as the upstream for AS149806. Hurricane Electric's peer table also shows AS132366 Alfaz Network as the observed IPv4 and IPv6 peer. Public route collectors are incomplete, and some private peering or national exchange arrangements may not be visible. Nonetheless, the public evidence supports a concentrated upstream structure. The company may hold an ASN, but it does not publicly display the redundancy of a multi-homed regional ISP with multiple transit providers and extensive peering.

ASN stewardship is an option, not independence

An autonomous system number is often confused with network independence. Economically, it is better understood as a property right bundled with an option contract. It gives the operator the ability to originate prefixes, choose upstreams, publish routing policy, build reputation, and move traffic without renumbering customers. These are real advantages over a pure reseller using upstream-assigned addresses.

For Speed Star, the option value is visible. The company controls AS149806 and originates its own IPv4 and IPv6 resources. Its public prefixes are RPKI-valid. Should an upstream contract become too expensive or unreliable, an operator with portable resources can, in principle, switch to another upstream or add a second while keeping the same addressing identity. This makes the ISP less fragile than a reseller whose customers are numbered from a wholesaler's block.

But an option is not the same as exercise. To exercise ASN autonomy, the operator needs practical alternatives: fiber paths to another upstream, compatible handover points, affordably priced ports, engineering capacity, routing policy, RPKI updates, billing terms, and regulatory authorization. In a local access market, these frictions can be significant. A second upstream may require new interconnections, new national transmission, new equipment, minimum commitments, and technical staff able to operate BGP during outages. The economic value of the ASN is therefore contingent on the upstream connectivity market.

This is why Speed Star is a good example of "waiting autonomy". It has the registry assets that make autonomy possible. Yet the visible routing evidence shows dependency on Alfaz Network. The network has a passport, but not necessarily many roads out of town.

The same logic applies to portable IP space. Portable addresses improve bargaining power because customers and NAT pools do not have to be renumbered when the upstream changes. But if domestic backhaul and gateway access are scarce or expensive, portability only partially reduces lock‑in. The bigger bargaining issue is not the address block; it is the physical and commercial path to the rest of the internet.

Retail demand: why small fixed-line ISPs survive alongside mobile giants

Bangladesh is an extremely mobile-internet-dominant market in subscriber numbers, but fixed broadband occupies a different use niche. AMTOB industry statistics, citing BTRC data, report 119.12 million mobile internet subscribers and 14.95 million ISP plus PSTN internet subscribers at the end of May 2026, for a total of 134.07 million internet subscribers. The same AMTOB page reports 188.60 million mobile subscribers. A Daily Star report on BTRC data indicates that fixed broadband connections were stable and gradually expanding, with 1.48 crore fixed connections in January 2026 after 1.46 crore between September and December 2025, even though mobile internet subscriptions declined over the preceding six months.

This matters because mobile and fixed are not perfect substitutes. Mobile data is personal, portable, and often capacity-limited by radio conditions and plan economics. Fixed broadband is home-based and premises-based. It supports shared video, telework, gaming, CCTV, Wi‑Fi calling, streaming boxes, and small-office use. A local fixed ISP does not need to beat mobile on every count. It needs to deliver a stable, cheap, unlimited or high‑capacity home connection in the buildings where its last‑mile fiber is already present.

The fixed broadband market in Bangladesh also remains fragmented. A BTRC broadband connectivity report references 2,715 ISPs and notes ISP plus PSTN user growth from 12.49 million in October 2023 to 13.74 million in October 2024, with 173,845 km of fiber optic deployment and a total network bandwidth of 6,600 Gbps. The same report identifies nationwide, divisional, district, and upazila/thana concession types, and references a 20 Mbps fixed broadband minimum standard.

Fragmentation creates room for firms like Speed Star. A neighborhood ISP can win by being physically close, cheap, and responsive. It can know which building managers allow cable access, which poles are congested, which customers pay on time, and which households will tolerate brief outages if field staff answer the phone. This local knowledge is hard for nationwide operators to replicate cheaply over the last few hundred meters.

But fragmentation also destroys pricing power. When there are many licensed ISPs and many informal or illegal operators, customers can compare simple monthly rates. A small access ISP cannot easily raise prices unless it offers superior reliability, real IP addresses, professional support, or building‑specific convenience. Speed Star's pricing grid fits this competitive logic: low advertised entry price, wide speed gaps, and premium IP features on higher tiers.

Revenue logic and margin pressure

The retail revenue model is simple: monthly subscription fees, installation charges where applicable, premium tiers, real IP add‑ons, business lines, and perhaps small ancillary services like router setup or dedicated bandwidth. The economics below the revenue line are more complex.

Assume a purely illustrative scenario. If Speed Star had 1,000 paying lines at 700 Tk average monthly revenue, gross monthly revenue would be 700,000 Tk. At 3,000 lines, it would be 2.1 million Tk. At 5,000 lines, it would be 3.5 million Tk. These numbers are not reported company financials; they are scale benchmarks. APNIC Labs customer population estimates appearing in search result snippets placed Speed Star in the several‑thousand‑user range at sampled dates in 2025 and 2026, but APNIC Labs measurements are not the same as billable subscriptions and can be affected by advertising measurement bias, NAT, device distribution, and sample composition.

The cost stack begins with upstream capacity. For a local ISP, international transit and domestic backhaul are not the same as retail Mbps. A 100 Mbps residential customer does not require 100 Mbps of guaranteed upstream capacity, but the ISP must buy enough peak bandwidth to keep congestion below the churn threshold. The operator also needs domestic reachability to BDIX and cache nodes. Local cache performance can substitute for some international transit, but only for content that can be cached and served locally.

Then come physical access costs: fiber drops, splitters, ONUs, routers, splicing, power, pole or duct arrangements, rooftop access, building permissions, repair inventory, and transport from the local access network to the upstream handover point. Small operators often underestimate these costs because much of the work is operational rather than capitalised. A field technician's scooter ride to fix a broken drop cable is not a glamorous infrastructure expense, but it is one of the main determinants of churn.

Customer support is another cost line. The website advertises 24/7 support and instant support. At a local ISP, support is not merely a service promise; it is a retention mechanism. A customer may tolerate an outage if a known person answers the phone and repairs the line quickly. Conversely, repeated unresponsiveness turns a broadband product with low switching costs into churn.

License and tax costs are not negligible. The ISPAB registers show a BTRC license number, a trade license, a BIN, and a TIN for Speed Star. The 2025 license transition raises the stakes because migration and renewal costs under the new FTSP categories can be significant for small operators. The provisional FTSP guidelines show revenue‑sharing and social‑obligation contributions, as well as application, acquisition, migration, annual, renewal, and bank guarantee amounts. For a district FTSP, the table includes an application fee of BDT 10,000, an acquisition fee of BDT 200,000, a migration fee of BDT 100,000, an annual fee of BDT 100,000, a renewal fee of BDT 200,000, a bank guarantee of BDT 200,000, a 5.5% revenue share, and a 1% social obligation contribution. For a full FTSP, the listed amounts are markedly higher.

These costs push the market toward a minimum efficient scale. A single‑neighborhood operator with low ARPU can survive if it has a dense customer base, cheap support, favorable upstream terms, and low bad debt. But the same operator becomes fragile if regulatory fixed costs rise, if upstream minimum commitments increase, or if a larger competitor enters the same buildings with promotional pricing.

The retail pricing environment intensifies this pressure. The BTRC's "One Country, One Rate" policy set indicative broadband pricing ranges in 2021, with reports describing 500 Tk per month for a minimum of 5 Mbps, 700 to 1,000 Tk for 10 Mbps per reports, and 1,100 to 1,200 Tk for 20 Mbps or more. The policy aimed to reduce geographic price discrimination and protect consumers, but press reports also noted operator complaints that IIG bandwidth prices and NTTN fees were high and implementation remained uneven.

For a firm like Speed Star, this creates a classic regulated‑retail, less‑regulated‑wholesale problem. Retail prices are constrained by policy, competition, and customer expectations. Wholesale and mid‑mile costs are shaped by supplier concentration, route geography, and contract terms. The margin is the residual. The local ISP holds the customer complaint relationship, but not all the inputs needed to control quality.

Upstream concentration and mid‑mile toll

The economically most important public fact about Speed Star's routing is its visible upstream concentration. BGP.tools shows AS149806 with a single upstream, Alfaz Network, and Hurricane Electric's peer table also identifies Alfaz as the observed IPv4 and IPv6 peer. Alfaz Network is itself a Bangladeshi AS with multiple upstreams visible in BGP.tools, including Summit Communications, Level3 Carrier and Rego Communications, and it appears to have downstream relationships that include Speed Star.

This implies layered dependency. Speed Star's immediate commercial and operational relationship may be with Alfaz. Alfaz's economics depend on larger upstream providers. Those larger providers depend on international capacity, national transmission, peering, content caches, and regulatory status. Each layer adds both service value and bargaining power extraction.

Mid‑mile providers have market power for five reasons. First, they aggregate demand from many small ISPs and can buy or peer at scales a local ISP cannot. Second, they control physical handover and national transport routes. Third, they have better access to content caches, IX ports, and international paths. Fourth, they can manage routing policy, filtering, RPKI, and capacity engineering at a professional scale. Fifth, regulatory rules have historically given particular license classes privileged roles in internet gateway and transmission functions.

Bangladesh's IIG regime made this visible. A Financial Express report indicated that thana and upazila‑level ISPs had been instructed to buy bandwidth from IIG providers from 1 July, with threats of regulatory action for non‑compliance. The same report described complaints that IIG bandwidth prices were high, IIG availability was not uniform across the country, long cable runs increased costs, and the IIGs themselves were not subject to the "one country, one rate" retail framework.

This is the structural squeeze. The customer sees Speed Star's 500 Tk or 800 Tk plan. Speed Star sees upstream bills, backhaul constraints, power cuts, equipment failures, and regulatory costs. A mid‑mile provider sees many Speed Stars and can set its prices based on the fact that the local ISP cannot easily tell all its customers to wait while it rebuilds its upstream path.

The 2025 license reform changes terminology but not necessarily the underlying economics. The Bangladesh Telecommunications Network and Licensing Policy 2025 states that legacy ICX, IIG, and IGW licences are to be phased out upon expiry. It introduces categories such as Access Network Service Provider (ANSP), National Infrastructure Connectivity Service Provider (NICSP), International Connectivity Service Provider (ICSP), and Fixed Telecommunications Service Provider (FTSP). The policy describes ANSPs as access network entities, NICSPs as infrastructure and transmission providers, and ICSPs as international connectivity providers. District FTSPs may provide internet and data services within a district and obtain international internet bandwidth from an ICSP or a FTSP.

For a small ISP, the key question is not whether the provider is called an IIG or an ICSP. The question is whether there are enough competing wholesale paths to discipline price and quality. If the reform increases contestability, small ISPs gain bargaining power. If it merely re‑labels chokepoints while raising compliance costs, consolidation pressure rises and mid‑mile market power persists.

Coronet as a benchmark for mid‑mile power

Coronet Corporation Limited is relevant because it shows the other side of the market. Its official website describes Coronet as an IIG and IP transit company in Bangladesh, offering MPLS, IPLC, Global Ethernet, dedicated internet access, and IP transit. Its PeeringDB profile identifies AS149765, says it operates as an IIG and as a nationwide retail ISP via another AS, reports traffic of 1–5 Tbps, and describes extensive peering with CDN/content providers and international internet exchange points. Hurricane Electric's page for AS149765 shows several internet exchange points, many observed BGP peers, and a far larger route footprint than Speed Star's.

This contrast is the market structure in miniature. Speed Star has hundreds of IPv4 addresses, a local address, a small set of originated prefixes, and one visible upstream. Coronet advertises service‑provider‑facing products, large traffic scale, extensive peering, and many observed peers. The small ISP sells to homes and small businesses. The mid‑mile provider sells to ISPs, mobile operators, and other service providers.

The difference in commercial power stems from scale. Coronet‑type operators can combine transit, peering, cache access, and national transport. They can handle large inbound content flows and negotiate with international operators. They can offer service‑level promises backed by multiple paths. Speed Star‑type operators can offer customer proximity and local installation speed, but they cannot easily replicate the upstream supply chain.

This does not mean that all small ISPs are economically doomed. The access edge has its own moat: local customer acquisition, last‑mile density, building relationships, fast repair, and trust. But it does mean that the surplus is contested. When retail prices are low and mid‑mile inputs are concentrated, the small ISP's margin depends on operational discipline and favorable procurement. The mid‑mile provider can capture a large share of the infrastructure surplus because it controls the scarce aggregation functions.

If Coronet appears alongside Speed Star in directories or routing datasets, the useful analytical inference is ecosystem proximity. Both are APNIC‑visible networks in Bangladesh; one is a major IIG/IP transit service provider, the other a small local access network. This proximity is economically sufficient to matter even without proof of ownership. It frames the bargaining environment in which Speed Star operates.

Route visibility: useful evidence with strict limits

BGP evidence is unusually valuable because it is hard to fake in the way marketing language can be. A network either originates prefixes visible to route collectors or it does not. Its ROAs are valid, invalid, or absent in observed validation systems. Its upstream paths are seen or not. For Speed Star, the public BGP evidence proves that AS149806 is an active routed network, not merely a dormant company registration.

But route visibility is not financial visibility. It does not reveal subscriber numbers, revenue, churn, uptime, support quality, private peering, precise fiber paths, building penetration, bad debt, or supplier pricing. Nor does it fully reveal domestic arrangements. A small ISP may have private BDIX connectivity, cache handoffs, or local transport relationships not visible in global route collector summaries. Conversely, a single visible upstream may understate operational relationships if traffic is exchanged privately.

The best use of BGP data is role classification. Speed Star looks like a consumer eyeball network because it originates a small address block, has no visible hosted domain base in IPinfo's dataset, advertises retail broadband plans, and displays consumer‑network behavioral tags. It does not look like a hosting provider, a major carrier, or a content network.

The second use is dependency analysis. Public collectors show a concentrated upstream path through Alfaz Network. This does not prove that no backup link exists, but it makes a robust multi‑homed architecture unlikely in the public internet view. If Speed Star added a visible second upstream, especially one that does not depend on the same domestic supplier chain, the economics would change. Redundancy would improve availability, bargaining leverage, and quality of service for higher‑value customers.

The third use is operational hygiene. Valid RPKI and consistent APNIC entities suggest that the network's registry state is maintained. This matters because the small‑ISP market in Bangladesh includes many operators with weak formal systems. Registry discipline is a modest but real signal of professionalization.

The fourth use is scarcity inference. The 512‑IPv4‑address pool, combined with retail broadband plans and APNIC Labs user population estimates in the several‑thousand range at sampled dates, suggests public IPv4 rationing and likely address sharing. This supports an economic reading of the real IP plans as a premium product. It does not establish exact CGNAT ratios or customer counts.

Customer churn: low formal switching costs, high local frictions

The economics of broadband churn in a dense urban neighborhood are paradoxical. On paper, switching is easy. Many ISPs may operate in the same locality; advertised plans are comparable; entry prices are low; and customers can call another provider. In practice, switching carries frictions.

The first friction is physical. A household may already have a fiber drop, an ONU, a router setup, and a cable path from Speed Star. Switching to another provider may require a new drop, a new installation appointment, new router settings, a new payment relationship, and possible downtime. In apartment buildings, available providers may be limited by roof access, building management, space in risers, or informal arrangements.

The second friction is service trust. A customer who knows a local technician or has a direct phone number may hesitate to switch to a cheaper but less responsive provider. Speed Star's public emphasis on 24/7 support and instant support is a direct response to this friction. In a low‑price market, service restoration often matters more than brand advertising.

The third friction is application‑specific. Customers with CCTV, gaming, remote access, small‑office VPN, or server‑type needs may depend on a public IPv4 address or stable NAT behaviour. Speed Star's higher tiers that include an IP address may reduce churn among these users. Once a customer has configured devices around a real IP, switching has higher setup costs.

The fourth friction is perceived performance. If BDIX, YouTube, and Facebook work well, many customers will not investigate the international routing table. They will judge the service by the applications they use most. Local caching and exchange connectivity can therefore reduce churn without forcing the operator to improve all destinations equally.

At the same time, buyer power remains real. Low‑income, price‑sensitive households may switch to save small monthly amounts. Students, renters, and small shops may churn when service breaks repeatedly. Mobile data remains a substitute for light users, while large national or regional ISPs may launch promotions. The local ISP's power is thus narrow: it can retain customers through convenience and support, but it has limited ability to raise prices above neighborhood norms.

Regulation: the hidden fixed cost of legitimacy

BTRC and ISPAB registers indicate that Speed Star's formal market position is that of a licensed local ISP, but current status requires caution. ISPAB gives the license number 14.32.0000.702.46.716.20.184 and says the license type is Upazila/Thana. The BTRC Upazila/Thana ISP list, dated 18 December 2024, includes Speed Star.Net at the Kuril/Vatara address with the same license number and a validity date of 13 February 2024 in the parsed row.

This is not enough to conclude that Speed Star was no longer licensed after that date. Public lists may be stale, renewal columns may be ambiguous, and an operator may remain listed while a renewal or migration is in progress. The active website, validated APNIC abuse contact, and active BGP visibility all suggest continued operation. But license‑status ambiguity is economically important. A small ISP's bargaining power with upstream providers, landlords, enterprise customers, and regulators depends heavily on formal compliance.

The 2025 licensing framework increases the importance of scale. Existing ISP categories are being migrated to the new FTSP categories, and the provisional guidelines state that existing ISP and PSTN licence holders must apply for migration to continue after expiry, with no financial incentive, subsidy, or exemption. The district FTSP category seems the most relevant path for an Upazila/Thana operator that wishes to remain local, while a full FTSP licence entails far higher fees and guarantees.

For Speed Star, the district FTSP economics are potentially significant. A small operator with a few thousand customers can absorb a migration fee of BDT 100,000 and an annual fee of BDT 100,000 more easily than a micro‑operator with a few hundred lines, but the 5.5% revenue share and 1% social obligation contribution are recurring levies on the margin. The bank guarantee also ties up capital. These costs create a regulatory economy of scale. They encourage consolidation, reseller relationships, or absorption by larger operators.

The policy also re‑draws vertical boundaries. The new framework distinguishes access, fixed telecommunications service, national infrastructure connectivity, and international connectivity. It allows district FTSP licensees to obtain international bandwidth from ICSPs or FTSPs, and permits fixed telecommunications service providers to build their own last‑mile connectivity, while NICSPs provide transmission infrastructure for lease and sharing.

The likely effect is not a simple deregulation story. It may reduce legacy inefficiencies, but it may also increase compliance burdens and strengthen large firms that can finance the guarantees, revenue share, and network obligations. Small ISPs that lack a dense customer base may become acquisition targets or wholesale‑dependent resellers. Operators with clean licenses, self‑owned ASN resources, and a stable local customer base will be better positioned.

Market power: who captures the surplus?

Speed Star's case shows that the broadband surplus is shared among at least six players: the retail ISP, the upstream aggregator, the national transmission provider, the international connectivity provider, the content/cache ecosystems, and the regulator. The household pays a single bill, but the revenue is economically pre‑allocated by the cost structure.

The retail ISP captures value through customer acquisition and last‑mile operations. It knows where the customers are, installs the line, collects payment, answers the phone, and absorbs complaints. This is a labor‑intensive, locally anchored activity. Its advantage is proximity.

The upstream aggregator captures value through scale. It aggregates many retail networks, buys larger capacities, manages BGP, maintains upstream diversity, and can connect to content‑rich paths. Alfaz Network's visible role as Speed Star's upstream illustrates this layer. It is smaller than a Coronet‑type IIG, but it still sits above Speed Star in the dependency chain.

The large IIG, ICSP, or transit provider captures value through even greater scale and route optionality. Coronet's public profile shows the economics of this layer: IP transit, dedicated internet access, MPLS, IPLC, Global Ethernet, extensive peering, and high traffic volumes. Such providers can sell reliability and reach to many smaller networks.

National infrastructure providers capture value through rights‑of‑way, fiber routes, ducts, poles, and backhaul. Even when international bandwidth becomes cheaper, the path from a neighborhood access network to a major aggregation point can remain expensive. Bangladeshi press reports on IIG dependency noted complaints about cable‑route costs and NTTN fees.

Content and cache providers capture value indirectly. They reduce the ISP's transit load and improve customer experience, but they also shape traffic concentration. If a few platforms dominate peak traffic, cache placement and peering access become quasi‑essential inputs. A local ISP advertising Facebook and YouTube caching is essentially telling its customers that it has optimised for the dominant applications.

The regulator captures value and imposes constraints through licensing, revenue sharing, social obligation contributions, pricing policy, and enforcement. The intent may be universal service, consumer protection, and an orderly market structure. The effect on small ISPs is a fixed‑cost hurdle and a compliance filter.

Supplier power is strongest where alternatives are scarce and switching is operationally risky. Buyer power is strongest where products are commoditized and physical switching is easy. Speed Star sits between the two. It faces customers with many options and suppliers with fewer options. That is why ASN stewardship matters: it is one of the few assets that can reduce supplier lock‑in.

Abuse, security, and reputation signals

The public abuse footprint shows no major negative signal. Scamalytics' ISP page for Speed Star Dot Net describes the network as low‑risk with a fraud score of zero in its dataset, while cautioning that its visibility is limited and its statements are opinions based on observed web traffic. CleanTalk's page for a Speed Star IP showed no current listing in its anti‑spam or SpamFirewall databases for that IP, while showing some ordinary network‑level spam‑rate and historical‑activity indicators. AbuseIPDB's AS page identifies AS149806, Bangladesh, and the two IP ranges corresponding to the 512‑IPv4 address block and the IPv6 space.

These signals should be weighted lightly. Residential eyeball networks regularly produce scattered abuse reports because customers use infected devices, misconfigured routers, spam software, VPNs, torrents, or compromised phones. IPinfo tags Speed Star with BitTorrent and VPN‑related signals as well as consumer ISP behaviour. This is not in itself a sign of a malicious provider. It is consistent with a small residential access network.

There was no credible public evidence in the reviewed material of a major outage, security breach, litigation, supply dispute, license cancellation, or systemic abuse problem involving Speed Star. Absence of evidence is not evidence of absence. Local service complaints may exist in Facebook groups, phone conversations, or neighbourhood channels that are not indexed in durable public records. For a company of this size, reputation is likely hyperlocal and may not leave a searchable trace.

The strongest risk signal is operational concentration. A single visible upstream creates outage correlation. A regulatory renewal gap creates sanction risk. A small address pool creates NAT and real IP support complexity. These are more economically important than the available abuse indicators.

Alternative hypotheses

The most likely hypothesis is that Speed Star Dot Net is an independent or locally controlled Dhaka access ISP, with its own ASN, portable IP resources, and upstream dependency on Alfaz Network. The supporting evidence is direct: APNIC registry records, ISPAB membership, the BTRC license list, the active website, and active BGP originations.

A second hypothesis is that Speed Star is a trading‑name operation or owner‑driven business whose formal legal wrapper is less visible than its network identity. This is common in small‑ISP markets. The evidence does not disclose directors or a company registry filing, and naming varies across sources. Economically, this would mean that management depth and funding capacity may depend on a small number of individuals rather than an institutional balance sheet.

A third hypothesis is that Speed Star is operationally captive to an upstream or mid‑mile provider, even if not legally owned by one. This would occur if the company's upstream contract, address administration, billing support, or field operations were effectively controlled by a larger partner. The visible single‑upstream relationship makes commercial dependence plausible, but it does not prove captivity. Economically, captivity would reduce Speed Star's independent bargaining power and convert part of its business into a local customer‑acquisition channel for the upstream ecosystem.

A fourth hypothesis is that the public BTRC license record is stale or incomplete against current operations. The December 2024 BTRC list includes Speed Star with a validity date in February 2024, while other records show ongoing website and BGP activity. This could reflect a renewal lag, list maintenance issues, migration status, or a compliance problem. Economically, a clean renewal or migration would support continuity; a true lapse would increase shutdown, disconnection, and acquisition risk.

A fifth hypothesis is that Coronet or another large provider is a direct counterparty. Coronet's public IIG/IP transit role and large peering footprint make it a plausible market actor around small ISPs, but the reviewed sources do not establish a direct relationship with Speed Star. If a direct Coronet relationship were proven, the economic interpretation would depend on its form. Transit supply would be ordinary wholesale dependence; reseller control would imply reduced independence; an acquisition would transform the firm from a standalone micro‑ISP into a retail edge of a larger network.

What the evidence proves, suggests, and leaves open

The evidence proves that Speed Star Dot Net is the APNIC‑registered identity associated with AS149806 and ORG-SSDN2-AP. It proves that the network originates a small IPv4 and IPv6 footprint visible in BGP, with RPKI‑valid routes in the reviewed public views. It proves that the operational brand speedstarbd.net advertises local broadband plans, BDIX connectivity, caching, fiber, support, and real IP features. It proves that ISPAB and BTRC registers tie a local Speed Star.Net ISP identity to Kuril/Vatara, Dhaka, and a specific Upazila/Thana ISP license number.

The evidence suggests that Speed Star is a small local fixed‑broadband access operator serving residential and small‑business users. It suggests a customer scale in the few thousands rather than hundreds of thousands, although public measurement estimates should not be mechanically converted into subscribers. It suggests likely IPv4 address sharing, because the public IPv4 pool is only 512 addresses while the service surface is a consumer broadband network. It suggests significant upstream bargaining exposure, because public collectors show a concentrated relationship via Alfaz Network.

The evidence leaves open the beneficial ownership of the firm, funding, directors, exact subscriber numbers, ARPU, churn, debt, supplier contracts, current license renewal or migration status, direct BDIX or cache topology, customer complaint history, fiber footprint, and any direct commercial relationship with Coronet. These unresolved facts are not decorative. Each would alter the economics.

If ownership is local and independent, the key risk is supplier dependency and regulatory fixed costs. If ownership is upstream‑connected, the key question becomes whether Speed Star is a margin‑generating retail business or a customer‑acquisition appendage. If subscriber numbers are closer to 500 than to 5,000, license migration costs could be severe. If closer to 5,000 dense lines, the firm may have a defendable local cash base. If the firm has a second private upstream not visible in public collectors, outage and bargaining risks are lower than the public routing table implies. If the BTRC license is not current, the risk profile changes immediately.

Infrastructure economics in a small network

Speed Star Dot Net reveals that the economics of small ISPs in Bangladesh do not reduce to bandwidth resale. The local ISP combines five functions: it is a neighbourhood sales organization, a last‑mile repair business, a billing collector, a manager of scarce addresses, and a buyer of mid‑mile reachability. Its competitive advantage is local density. Its vulnerability is upstream dependence.

ASN stewardship gives the firm a formal network identity. That identity has option value because it can survive upstream changes and support real IP monetisation. But ASN stewardship does not eliminate market power above the ISP. If the operator has only one practical upstream path, the supplier can still capture much of the surplus. If retail tariffs are constrained and customers are price‑sensitive, the ISP cannot pass on every upstream cost increase. If regulatory migration adds fixed costs, small scale becomes a handicap.

The firm also shows why route visibility is both important and misleading. Public BGP data can identify network role, address scarcity, upstream concentration, and route hygiene. It cannot say whether the field technician arrives in 20 minutes, whether customers pay on time, whether building managers favour an ISP, or whether the upstream contract is punitive. Infrastructure intelligence must therefore combine routing data with licenses, tariffs, market structure, and local service evidence.

The most important conclusion is that mid‑mile providers hold disproportionate market power because they sell the scarce input that converts local access into internet service. A small ISP may hold the customer relationship while lacking control over quality and costs. In this sense, Speed Star is not only a company. It is a measurement point in the broadband political economy of Bangladesh: local retail competition at the bottom, concentrated routing and transmission dependencies above, and a regulatory transition pressing from the side.

Evidence register

1. APNIC/RDAP initial registration for AS149806, Speed Star Dot Net, entity handle ORG-SSDN2-AP:https://rdap.org/autnum/149806. This is the target identification anchor provided for the report; APNIC WHOIS renderings provide the corroborative registry fields used above.
2. APNIC WHOIS, inet6num 2400:7520::/32, SPEEDSTARNET-BD, Speed Star Dot Net, ORG-SSDN2-AP, KA-119/A, Kuril, Dhaka 1229, portable allocated IPv6 space, validated abuse contact:https://wq.apnic.net/apnic-bin/whois.pl?object_type=inet6num&searchtext=2400%3A7520%3A%3A%2F32.
3. APNIC WHOIS, organization entity ORG-SSDN2-AP, Speed Star Dot Net, Bangladesh LIR identity, phone and email fields:https://wq.apnic.net/apnic-bin/whois.pl?form_type=advanced&searchtext=ORG-SSDN2-AP.
4. Hurricane Electric BGP Toolkit, AS149806, Speed Star Dot Net, APNIC aut-num rendering, originated prefixes, RPKI status, observed peer Alfaz Network:https://bgp.he.net/AS149806.
5. BGP.tools, AS149806, Speed Star Dot Net, active Bangladeshi eyeball network, upstream Alfaz Network, originated prefixes and RPKI-valid status:https://bgp.tools/as/149806.
6. IPinfo, AS149806, Speed Star Dot Net, Bangladeshi ISP, 512 IPv4 addresses, consumer ISP and stub‑AS signals:https://ipinfo.io/AS149806.
7. Speed Star.Net official website, service claims, BDIX connectivity, caching, dedicated IP, fiber, pricing grid, contact details:https://www.speedstarbd.net/.
8. ISP Association of Bangladesh member profile for Speed Star. Net, membership A-332, BTRC Upazila/Thana license number, establishment date, trade license, BIN, TIN, website and contact data:https://ispab.org/member/speed-star-net.
9. ISP Association of Bangladesh member list entry for Speed Star. Net, license category and Kuril/Vatara address:https://ispab.org/members/S?page=5.
10. BTRC Upazila/Thana ISP License List as on 18 December 2024, including Speed Star.Net at Ka-119/1, Kuril, Vatara, Dhaka-1229, license number 14.32.0000.702.46.716.20.184:https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/idjz6hji0u8v/b/btrc-prod/o/static%2Fuploaded%2Flicense%2FISP%28Upazila_Thana%29%20License%20List%20as%20on%2018-12-2024%20%281%29%20%281%29%20%281%29_29e9f4bf494145f5bfee76bd1a384ddc.pdf.
11. BGP.tools, AS132366 Alfaz Network, visible upstreams, peers, downstream relationship context including Speed Star:https://bgp.tools/as/132366.
12. Hurricane Electric BGP Toolkit, AS132366 Alfaz Network, RPKI and peer observations:https://bgp.he.net/AS132366.
13. Coronet Corporation Limited official website, IIG and IP transit service descriptions:https://coronetbdiig.com/.
14. PeeringDB, Coronet Corporation Limited, AS149765, network type, traffic scale, peering policy, IIG and nationwide retail ISP description:https://www.peeringdb.com/net/32178.
15. Hurricane Electric BGP Toolkit, AS149765 Coronet Corporation Limited, internet exchange points, prefixes, peers and route footprint:https://bgp.he.net/AS149765.
16. APNIC Labs customer population estimates for Bangladesh by AS, including sampled estimates for Speed Star and Coronet; useful as a directional measure, not as subscriber accounting:https://stats.labs.apnic.net/cgi-bin/aspop?c=BD&d=07%2F08%2F2025and associated date views.
17. AMTOB industry statistics citing BTRC data, May 2026, mobile subscribers, mobile internet, ISP plus PSTN and internet subscriber totals:https://www.amtob.org.bd/home/industrystatics.
18. Daily Star, internet subscriber base and fixed broadband stability, January 2026 BTRC subscriber figures:https://www.thedailystar.net/business/economy/news/internet-subscriber-base-shrinks-70-lakh-6-months-4128611.
19. BTRC broadband connectivity report, ISP concession types, fixed broadband minimum speed standard, ISP count, fiber deployment and fixed broadband growth context:https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/idjz6hji0u8v/b/btrc-prod/o/report%2F2553c9a48743467faaa8b420c2e6ecb5.pdf.
20. Daily Star, "One Country, One Rate" broadband pricing framework and BTRC tariff ranges:https://www.thedailystar.net/toggle/news/one-country-one-rate-internet-flat-rate-across-the-country-soon-2105929.
21. The Business Standard, rural broadband users, tariff implementation, illegal ISP concerns and operator complaints about IIG and NTTN costs:https://www.tbsnews.net/bangladesh/telecom/broadband-internet-rural-users-still-pay-higher-get-lesser-speed-277135.
22. Financial Express, BTRC directive requiring zonal or thana/upazila ISPs to buy bandwidth from IIG providers and industry complaints about cost and quality:https://thefinancialexpress.com.bd/home/zonal-isps-have-to-buy-bandwidth-from-iig-service-providers-1650680035.
23. Bangladesh Telecommunications Network and Licensing Policy 2025, phasing out of legacy IIG/IGW/ICX categories and introduction of ANSP, ICSP, NICSP and related categories:https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/idjz6hji0u8v/b/btrc-prod/o/notice%2Fdc545fba9f8b4a39851a4f3290f5573c.pdf.
24. Draft Regulatory and Licensing Guidelines for Fixed Telecommunications Service Providers, migration rules, renewal conditions, fees, bank guarantees, revenue sharing and social obligation contributions:https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/idjz6hji0u8v/b/btrc-prod/o/notice%2Fed7fe810835b43a8ac5243cfd782bfce.pdf.
25. Scamalytics, Speed Star Dot Net ISP risk page, low fraud score with explicit limitation caveat:https://scamalytics.com/ip/isp/speed-star-dot-net.
26. CleanTalk, sample Speed Star IP and network‑level anti‑spam reputation signals:https://cleantalk.org/blacklists/103.186.216.92.
27. AbuseIPDB, AS149806 page, IP range identification and Bangladeshi AS context:https://www.abuseipdb.com/check/AS149806.
28. Hurricane Electric country page for Bangladesh networks, useful as a comparative routing directory for Bangladeshi ASNs, including small access networks and large providers:https://bgp.he.net/country/BD.

Watchpoints

The first watchpoint is license migration. Confirmation that Speed Star has renewed, migrated, or been reclassified under a district FTSP license or another new category would materially reduce regulatory ambiguity. Absence of migration would increase the probability of upstream disconnection, enforcement action, forced resale, or absorption by a larger operator.

The second watchpoint is visible upstream diversification. A second independent upstream in BGP, especially one that does not rely on the same national aggregation chain, would alter the firm's economics by improving availability, reducing supplier lock‑in, and giving Speed Star more credible procurement leverage.

The third watchpoint is a change in the relationship with Alfaz. If AS149806 disappears from Alfaz's downstream view, if Alfaz changes its own upstream mix, or if Speed Star shifts to a larger Coronet‑type IIG or ICSP, the mid‑mile cost and reliability model should be re‑estimated.

The fourth watchpoint is RPKI and prefix stability. A route withdrawal, invalid ROA, new IPv4 allocation, additional IPv6 de‑aggregation, or originating AS change would signal a significant operational or control event. New address resources would also alter the real IP monetisation economics.

The fifth watchpoint is subscriber density evidence. APNIC Labs population estimates, local advertising intensity, plan changes, installer hiring, payment‑agent expansion, or neighbourhood customer comments can reveal whether the operator is a dense cash‑generating local business or a thin reseller of limited scale.

The sixth watchpoint is real IP pricing. Any change in plans that makes public IPv4 cheaper, scarcer, or separately charged would indicate shifting address pressure, customer mix change, or upstream sourcing. IPv4 scarcity is one of the few levers through which a small ISP can create premium ARPU.

The seventh watchpoint is FTSP, ICSP, and NICSP implementation. If the new licensing regime increases wholesale competition, Speed Star‑type operators gain bargaining power. If it raises fixed costs without increasing supplier contestability, consolidation pressure will rise.

The eighth watchpoint is service quality evidence. Repeated local outage complaints, slow‑repair reports, payment disputes, or support failures would matter more than generic abuse telemetry, because churn in a neighbourhood ISP is determined by experienced quality of service.

The ninth watchpoint is disclosure of acquisition or control. Any filing, association update, domain transfer, upstream‑branded plan, or address change linking Speed Star to a larger ISP, IIG, or infrastructure provider would transform the analysis from an independent local ISP economics to a retail‑edge consolidation economics.

The tenth watchpoint is fixed‑wireless and mobile substitution in Kuril/Vatara. If 5G fixed wireless, improved mobile plans, or alternative fiber entrants become significantly better or cheaper in the same buildings, Speed Star's local switching frictions will weaken and its dependence on quality of service will increase.