The Exchange That Should Have Made Bangladesh Cheaper

A company whose name promises an institution

“Bangladesh Internet Exchange Ltd” sounds like the sort of company that should sit at the center of a national cost revolution. In any under-peered market, a true exchange lowers the marginal cost of domestic traffic by making local routes cheaper than international detours. It also shortens path length, improves latency, and weakens the bargaining power of upstream transit sellers. In Bangladesh, where regulators explicitly framed National Internet Exchanges as instruments to keep domestic traffic inside the country and save foreign currency, a company with that name ought to have mattered a great deal. The commercial puzzle is that the current public record does not primarily describe a visible, neutral, traffic-localization institution. It describes something thinner and stranger: a legal and registry-visible Bangladeshi telecom company with historical wireless-broadband ambitions, dormant-looking public routing traces, asset-transfer signals, and only fragmentary evidence of current exchange-like operations.

That distinction matters because an Internet exchange is not economically real merely because the words “internet exchange” appear in a company name or registry entry. It becomes economically real when there is enough visible interconnection for other operators to trust a new local price for reachability. In practice that means public member disclosure, route-server usage, visible peering LANs, disclosed PoPs, traffic statistics, operational contacts, and governance that convinces competitors to cooperate on one layer while competing on others. Bangladesh’s own NIX rules anticipated exactly this: they define a NIX as the exchange point for domestic traffic, define a Looking Glass server as a route collector, require quarterly reporting including active users and bandwidth utilization, and require traffic data to be maintained on the licensee’s website. The rules were written for institutions that make routing and volume visible enough for a market to coordinate around them.

On that standard, Bangladesh Internet Exchange Ltd does not presently look like the exchange that made Bangladesh cheaper. The strongest current public proof of existence is not an exchange website, public port list, traffic graph, or PeeringDB exchange record. It is an ISPAB member page identifying the company as a Bangladeshi “Nationwide” license holder, but with no public directors list, no published website, no address list, no PoP list, no BTRC license number, and no trade-license, BIN, or TIN details filled in. Even the page’s own text invites the company to complete the missing verification details. That is a public identity trace, not a public exchange institution.

The company’s old commercial traces reinforce the ambiguity. Older directories and trade listings describe Bangladesh Internet Exchange Ltd, often branded “BD Net,” as an ISP with Banani addresses and a DNS Group website, and explicitly sell the company as a way to handle local mail and web traffic locally to avoid costly international links. A 2004 trade-press reference places Bangladesh Internet Exchange Ltd inside “DNS Group” alongside BDTEL Communication and DNS Software. In other words, the original pitch was economically coherent: localize traffic, save international bandwidth, and own the domestic choke point. But those old listings now read less like proof of a functioning exchange and more like a fossil of an unrealized or later-diverted strategy.

The reason this matters is simple. If the company had become a genuine domestic peering utility early enough, Bangladesh should have experienced a stronger local peering equilibrium earlier than it did. The country’s policy framework wanted that outcome. The economics rewarded that outcome. Yet the public market structure that emerged places the center of visible peering elsewhere, mainly in BDIX and later licensed competitors such as ISPAB-NIX, rather than in Bangladesh Internet Exchange Ltd itself. The company’s name therefore captures a lost possibility more than a proven outcome.

What the public record actually shows

The safest starting point is identity. Public RIR-related evidence ties the company to Bangladesh. APNIC’s transfer logs explicitly identify “Bangladesh Internet Exchange Ltd” with country code BD, and public ASN mirrors identify AS37994 as “BIE-AS-BD” registered to Bangladesh Internet Exchange Ltd under APNIC. A second ASN, AS56115 “NGGL-BD,” is also publicly associated with Bangladesh Internet Exchange Ltd in current mirror data. Those records do not prove active service, but they do strongly support that this is a real Bangladeshi telecom-resource holder and not a mistaken lookup of the better-known BDIX trust.

What they also show, however, is dormancy in the visible routing layer. IPinfo and DB-IP both describe AS37994 as having zero visible IPv4 and zero visible IPv6 ranges now, and DB-IP shows the same zero-prefix condition for AS56115 through its “other ASN allocated” listing. Public mirrors of AS56115 likewise show no IP ranges attached. These are not dispositive measurements of all private activity, but they are commercially meaningful: a company that is actively functioning as a nationally relevant exchange or route-localization utility usually leaves a stronger public routing exhaust than zero visible prefixes on its named ASNs.

There are historical routing traces, but they mostly read like earlier network life rather than present exchange vitality. The APNIC-derived WHOIS shown by bgp.tools for Bangladesh Bank’s AS45532 still contains an import and export policy with AS37994, meaning Bangladesh Internet Exchange Ltd once appeared in at least some formal interconnection policy as a routing counterparty. That matters because it proves the company was not purely a paper registry artifact. At some point it was integrated enough into local routing relationships that another Bangladeshi network encoded policy around it. But it does not prove anything about present scale, current port traffic, or whether those relationships still carry meaningful volume.

The company’s public corporate and technical contacts also suggest continuity mixed with opacity. Older listings name Rafel Kabir as managing director and associate the business with Banani offices, dnsgroup.com.bd, and bdtel.net. Current APNIC mirror-style pages still expose Rafel Kabir-era contact details in historical WHOIS fragments while also showing an organization email of lir@multinet.net, which points to a later Multinet-linked period. Meanwhile, the present ISPAB page surfaces a different public-facing contact, “Mr. Mihir Kanti Puri,” plus an email and mobile number, but no directors, no verified corporate website, and no visible facility footprint. The effect is not that the company disappears. The effect is that ownership/control history looks layered and only partly public.

That ownership layering becomes clearer when the historical wireless-broadband story is added. Multiple local press reports from 2013 and 2014 identify Bangladesh Internet Exchange Ltd as the operating company behind OLLO, first as a fixed WiMAX provider, then as the recipient of a third BWA/LTE-related license and 40 MHz of 2.6 GHz spectrum, and as one of two Ollo-linked entities alongside New Generation Graphics Ltd. The Daily Star called BIEL a subsidiary of Multinet Group in 2013, bdNOG speaker biographies later described Ollo as a Russian-owned multinational operating in Bangladesh under the BIEL name, and DSL’s “Sister Concerns” page described Bangladesh Internet Exchange Ltd as a joint venture with Multinet, Cyprus. So the company became, at minimum, more than an exchange. It became a wireless-access and spectrum story.

That pivot is crucial. The moment a would-be exchange operator becomes a spectrum operator and retail access aspirant, its incentives change. A neutral market-maker wants as many rival networks as possible to meet on fair terms. A retail ISP or wireless operator, by contrast, may value asymmetric control, captive customers, and privileged access to scarce regulatory licenses. The public record around Ollo shows precisely those tensions. Dhaka Tribune reported in 2014 that BIEL and NGGL sought to merge their spectrum, that the arrangement raised legal objections because telecom licenses and assigned frequencies were not meant to be transferable, and that market insiders viewed the combined spectrum valuation as unusually generous. Whether one accepts the rhetoric or not, the economic point is the same: this history looks like contested regulatory arbitrage in wireless access, not the clean development of a neutral exchange institution.

The official BTRC annual report confirms at least the regulatory skeleton of that story. In its 2018-2019 report, the regulator says Bangladesh Internet Exchange Ltd received the third Broadband Wireless Access license in 2013 with 40 MHz in the 2.6 GHz band. The same report also lists Bangladesh Internet Exchange Ltd among wireless ISPs with assignments in both the 800 MHz and 3.5 GHz ranges. Officially, then, the company was more than a NIX aspirant; it held wireless-spectrum resources, and BTRC still described BWA as intended to carry high-speed data and voice into underserved areas. That confirms the mixed identity problem at the heart of this case.

Yet the regulator’s own NIX framework was built around a different governance logic. The 2012 NIX guidelines say existing NIX providers had to obtain a NIX license and be certified by at least ten licensed ISPs with peering connectivity to the applicant. They also say a consortium of two or more licensed ISPs seeking a NIX license had to form a new joint-venture company. In other words, the rulebook imagined exchanges as shared infrastructure with broad operator buy-in. That is almost the opposite of a vertically tilted wireless operator strategy. Bangladesh Internet Exchange Ltd may have had “exchange” in the name, but the available evidence suggests its history drifted toward spectrum and ISP economics rather than toward trusted exchange governance.

The newest public signals add a final twist. APNIC’s live transfer log shows Bangladesh Internet Exchange Ltd as the source organization for a transfer of 203.188.160.0/19 to a RIPE NCC recipient on 9 September 2024 and another transfer of 27.0.96.0/19 to Byteplus Pte. Ltd. on 23 October 2024. APNIC’s transfer-log page snippet also shows a 2014 M&A ASN transfer involving ASN131770 to Bangladesh Internet Exchange Ltd from an Indonesian source. None of this proves distress, retreat, or a specific motive. But it does prove that BIEL has been active in the tradable telecom-resource layer. Economically that matters because a company whose public edge has faded but whose resource transactions remain visible may be functioning more as a resource-and-license vehicle than as a transparent exchange utility.

One slim current-service clue points in another direction, and it deserves to be stated carefully. Bangladesh Submarine Cables PLC’s 2024 annual report includes a customer line item labeled “Bangladesh Internet Exchange Ltd-IIG.” That is not enough to prove full current IIG operating status, traffic scale, or profitability. But it is enough to say that public commercial records still place the company somewhere in the gateway/bandwidth customer map rather than purely in historical memory. The company is therefore not best understood as fictional. It is better understood as a still-existing telecom entity whose current exchange role is unproven and whose clearer public activity is in licenses, resources, and residual operator traces.

Why Bangladesh still pays for domestic leakage

To answer the central economic question, it helps to state the mechanism plainly. A well-run exchange lowers national Internet cost in three stages. First, it changes the routing default: local destinations become reachable over local peering rather than via international transit. Second, it changes bargaining: if enough traffic can be exchanged locally, upstream transit vendors lose some leverage over domestic route pricing and quality. Third, it changes gravity: CDNs, caches, DNS anycast nodes, and local hosting follow the concentration of eyeballs and interconnection, which compounds the savings. A “registry-visible exchange” that is visible only in legal or ASN databases, but not in member participation and traffic, does none of that. It cannot discover a domestic price because no one can see or trust the market.

Bangladesh as a country still shows the symptoms of incomplete local price discovery. Internet Society Pulse says Bangladesh had eight active IXPs with 178 combined members as of June 2026, yet IXPs were present in only one of the country’s seventeen population centers above 300,000 people. APNIC’s 2025 Bangladesh peering article goes further and says only about 7.57% of registered networks are peering at NIXs, even though NIX capacity and membership have been growing. Those two facts are not contradictory. They mean Bangladesh has a visible exchange layer, but it is concentrated, incomplete, and still far from universal direct interconnection. In such a structure, a lot of domestic traffic still does not face an obvious, trusted local route first.

The physical geography of the exchange market reinforces the problem. BDIX, the country’s first and largest IXP, has a real public footprint: PeeringDB describes it as not-for-profit, open, and neutral, lists multiple PoP locations and LAN prefixes, and ISOC Pulse counts 147 members, 77 of them using the route server. That is exactly what public exchange visibility should look like. ISPAB-NIX also has a clear PeeringDB footprint, public IX-F export URL, public contact details, and a stated operating license date in 2020. KTL-IX and Summit NIX have similarly public records and explicit domestic-latency/value propositions. The price-discovery layer in Bangladesh is therefore no longer absent. It is occupied by other, more visible exchanges. That is one reason Bangladesh Internet Exchange Ltd matters less than its name implies.

But even the visible exchange layer does not yet dominate national traffic economics. Bangladesh’s larger cost problem is that the ISP market is broad, fragmented, and transit-sensitive. ISOC Pulse’s country report gives Bangladesh an “excellent” ISP-choice score and says 53% of the population used the Internet in 2024, while APNIC’s resilience commentary notes 34 registered IIGs as of mid-2023 and more than 2,000 hyper-local ISPs. A fragmented retail edge often creates an awkward peering equilibrium: many small operators care about minimizing monthly cash outlay more than building sophisticated bilateral peering, and they often buy cheap transit from intermediaries rather than invest engineering effort into full local interconnection. The exchange can offer savings, but the savings are unevenly distributed and require coordination.

The outage record shows how exposed that structure remains. During the October 2023 Khawaja Tower fire in Dhaka, Internet connectivity for around 40% of users was disrupted, showing how much of the country’s telecom flow still clustered around a few facilities hosting gateways, data centers, and interconnection equipment. ISOC and APNIC both used the incident to argue for greater resilience, more distributed interconnection, and less concentration of “Internet-enabling infrastructure.” In a more mature domestic exchange economy, internal traffic localization and distributed facilities would cushion that kind of shock better. The episode showed that Bangladesh had improved significantly, but it had not yet converted local interconnection into enough structural autonomy.

Submarine-cable dependence tells the same story from another angle. When SEA-ME-WE 5 suffered an outage in April 2024, Internet Society reported a roughly 25% increase in latency for services usually reached from Singapore. Bangladesh had backup through SEA-ME-WE 4 and terrestrial links with India, but the event highlighted how much of the country’s digital economy still depends on foreign path options and foreign hubs. This is where a working domestic exchange should help most with local traffic: it cannot remove the need for international transit, but it can sharply reduce the amount of traffic that unnecessarily touches it. The fact that international-path shocks still map quickly into user experience means localization is improving, yet not deeply enough.

There is an important subtlety, however. Not all “localization” has to happen at an exchange. In Bangladesh, local cache servers now carry much of the economic burden. Internet Society says around 69% to 70% of Bangladesh’s top 1,000 sites are available from in-country servers or caches, and The Business Standard reported during the July 2024 shutdown period that more than 80% of Bangladeshi traffic had normally been served by local caches. Without those caches, international bandwidth consumption would have risen five to six times, and broadband speeds collapsed. That means the largest immediate savings in Bangladesh are increasingly generated not by domestic inter-ISP traffic alone, but by global-content firms and caching partners sitting close to eyeballs. For an exchange, this changes the game. The exchange must attract CDN gravity or offer efficient access to caches, not merely connect local ISPs to one another.

This is precisely why the question of Bangladesh Internet Exchange Ltd is so sharp. If the company had become the obvious neutral platform where local ISPs, eyeball networks, banks, government networks, and cache operators met, it could have created a strong shadow price for domestic reachability. Instead, the public evidence suggests that the company migrated toward access-network and resource economics, while the actual peering gravity pooled around BDIX and later around competing licensed NIXs. So Bangladesh’s domestic leakage problem was addressed, but by an ecosystem rather than by the named company that seemed built for it.

Where price discovery happens now

Price discovery in peering markets is almost never a quoted price board. It is encoded in who joins, who uses the route server, where caches land, and whether smaller networks believe that local connectivity is reliable enough to substitute for upstream transit. In Bangladesh today, the clearest public venue for that discovery is BDIX. It has the largest publicly visible member count, multiple facilities, published traffic statistics, open/neutral marketing, and enough member diversity that the exchange itself becomes a reference point. ISOC Pulse says 147 ASNs are listed there; PeeringDB shows several Dhaka and nearby facilities; and the exchange’s public profile includes both IPv4 and IPv6 LAN prefixes. Those features are what turn a switch fabric into a market.

ISPAB-NIX has become the second important venue because it lowers onboarding friction. Its PeeringDB entry states a BTRC operating license date of 7 September 2020 and, unlike BDIX, publishes an IX-F member export URL openly. Public export matters because it reduces search costs. Potential peers and researchers can inspect membership and route-server use more easily, which lowers information asymmetry. Economically, that is not cosmetic. A market with lower search costs and more legible counterparties discovers better peering outcomes because operators can compare alternatives instead of buying transit by default.

The newer licensed exchanges compete by specializing around latency, geography, and operator bundles. KTL-IX explicitly pitches itself as a way to avoid dragging domestic traffic “all the way to US/SG/Abroad,” and Summit NIX sells itself on the strength of Summit Communications’ broader IIG and transmission footprint. Those are classic Bangladesh-specific peering propositions. They say, in effect: if we can combine domestic switching with a nationwide backbone or gateway relationship, we can arbitrage the country’s geography and lower both path length and coordination cost. Whether they succeed at national scale remains uneven, but the public pitch itself confirms that the controlling economic variable in Bangladesh is still transit avoidance.

Yet even in these visible exchanges, bargaining dependence is not fully broken. ISOC Pulse says Bangladesh’s domestic network coverage at IXPs is 59% when member networks and their customer cones are counted, but APNIC’s Bangladesh peering analysis still says only around 7.57% of registered networks directly peer at NIXs. Meanwhile BDIX route-server participation is only 77 of 147 members. That gap matters. Direct peering is where the strongest market discipline usually forms. If a large share of networks remain outside direct peering or avoid the route server, then domestic traffic pricing still depends on bilateral negotiations, selective peering, and the fallback option of paid transit. The exchange exists, but the exchange does not fully clear the market.

There is also a route-trust problem hiding inside the market design. APNIC’s Bangladesh analysis says the country’s ROA coverage is very high, around 98% for IPv4 and 96% for IPv6, but Route Origin Validation adoption is under 1%. ISOC Pulse says BDIX does not participate in MANRS and does not host a RIPE Atlas anchor. Those facts do not mean Bangladesh’s exchanges are unsafe. But they do mean the market is still incomplete in one of the most commercially important ways: operators can register intent and sign up as members, yet the collective routing-security posture remains weak enough to preserve caution. Where route trust is weak, selective peering and upstream dependence persist longer than pure cost models would predict.

That helps explain a persistent Bangladesh pattern noted in operator discussions. APNIC’s 2023 South Asia panel summary says a key reason local IXP traffic was not growing faster was that much OTT content was not local, while national long-distance route capacity also constrained intra-country transport. That is a subtle but decisive point. An exchange can only discover low domestic prices if the whole chain is efficient: metropolitan exchange ports, long-haul domestic transport, data-center gravity, and CDN/cache siting. If one of those links is weak, traffic still prefers foreign or upstream paths even when a domestic exchange exists on paper. Bangladesh’s constraint, then, is not only “more IXPs.” It is whether direct local peering, domestic backhaul, and content placement reinforce one another strongly enough to beat the default transit path.

In that sense Bangladesh Internet Exchange Ltd’s weak current public footprint matters symbolically as much as commercially. The name suggests a company that could have become a national reference venue for domestic route value. Instead, the market’s visible reference venues are other institutions. Bangladesh therefore did get exchange-led price discovery, but less concentrated, less universal, and later than such a name would have promised.

Why Bangladesh Internet Exchange Ltd never became that market-maker

The first reason is governance mismatch. Bangladesh’s own NIX architecture was designed around coalition trust among licensed ISPs. The BTRC guidelines require evidence of peering from at least ten licensed ISPs and contemplate consortium-based joint ventures. A neutral exchange in such a framework is a club good: competitors use it because none of them fully controls it. Bangladesh Internet Exchange Ltd’s public history, by contrast, runs through private ISP branding, DNS Group lineage, Multinet-linked control, WiMAX/LTE licensing, and Ollo spectrum aggregation. That trajectory may have been strategically rational for investors seeking scarce licenses. It was not the cleanest path to becoming a trusted neutral interconnection commons.

The second reason is that access economics overwhelmed exchange economics. BIEL’s major visible regulatory wins were in broadband wireless access and LTE. That placed management attention and capital demand in a completely different part of the value chain. Wireless access businesses burn money on spectrum, radios, customer acquisition, and coverage. Exchanges, by contrast, win by compounding neutrality, member density, and operational transparency. Once BIEL became publicly associated with Ollo’s WiMAX/LTE path, the company’s strategic identity shifted toward being a broadband operator with exchange-like history, not an exchange with broadband-adjacent services. In markets with limited management bandwidth and volatile regulation, that type of shift is often fatal to the exchange thesis.

The third reason is that public visibility itself decayed. BTRC’s NIX rules expect route-collector visibility and website traffic reporting. By contrast, Bangladesh Internet Exchange Ltd’s current ISPAB page shows no public website, no public PoP list, no public directors list, and no public license number. Its ASNs show no current visible prefixes in common public mirrors. That combination is devastating to exchange economics. Operators do not treat an exchange as a normal market venue if they cannot inspect the venue. Opacity raises perceived counterparty risk and raises the cost of technical due diligence, both of which push smaller networks back toward upstream transit or better-known IXPs.

The fourth reason is that the biggest bandwidth savings in Bangladesh migrated toward cache economics rather than pure domestic peering. Once Google, Meta, Akamai, Cloudflare, TikTok/Bison and similar content systems place caches locally, a large share of the international-bandwidth bill falls even without a specific exchange operator dominating the market. During the 2024 shutdown-cache disruptions, operators said more than 80% of normal traffic had been served locally by cache servers and that losing those caches dramatically increased international-bandwidth dependence. In a world like that, a would-be exchange no longer wins simply by saying “keep local traffic local.” It must also become the natural aggregator or access point for cache gravity. Public evidence does not show Bangladesh Internet Exchange Ltd establishing that role.

The fifth reason is that Bangladesh’s actual exchange market already found a first mover and later a set of visible followers. BDIX’s long history, public branding as a not-for-profit neutral exchange, facility distribution, and route-server base made it the obvious coordination point. APNIC’s community material still refers to BDIX as Bangladesh’s first and largest exchange, and public biographies of Bangladeshi network leaders repeatedly tie the country’s interconnection development to BDIX rather than to BIEL. Once that equilibrium forms, it is hard for another entity—especially one with a mixed retail-spectrum identity—to displace it. Exchange markets exhibit network effects. The members already where the members are tend to keep winning.

The sixth reason is that BIEL’s publicly visible asset behavior now looks more like portfolio management than exchange compounding. APNIC’s 2024 transfer records show two sizable IPv4 blocks leaving the company, including a transfer to Byteplus. A resource sale is not automatically negative; in telecom, it may reflect restructuring, monetization, or simple non-core optimization. But exchanges that are trying to become national liquidity venues usually market growing interconnection richness, not shrinking visible address inventory and dormant public BGP presence. The observed pattern is more consistent with a company preserving, reallocating, or monetizing legacy telecom assets than with one scaling a public peering marketplace.

The final reason is more structural and less company-specific: Bangladesh’s broader market still rewards transit intermediation. The country has many IIGs, multiple terrestrial cable ties to India, relatively concentrated data-center gravity in Dhaka, and thousands of smaller ISPs. In that environment, a truly neutral exchange can lower costs, but it must work against an entrenched commercial habit: many smaller networks buy convenience, not just bandwidth. They prefer a transit seller or gateway operator who can simplify international and domestic reach together. That leaves less room for a thin, purely domestic exchange thesis—especially if the exchange operator itself is not visibly superior at openness and trust.

So the answer to the core question is conditional. Yes, in principle, a registry-visible exchange can lower domestic transit leakage, latency, and bargaining dependence in Bangladesh. But only if registry visibility is matched by operational visibility, trusted governance, enough member density to become the default local path, and enough content/cache gravity to matter in real traffic terms. Bangladesh Internet Exchange Ltd does not currently satisfy those conditions in public evidence. Its current public role is too thin, too mixed, and too opaque.

What the remaining signals still prove

It would be wrong, however, to conclude that Bangladesh Internet Exchange Ltd is merely a ghost. The public record proves more than that. First, it proves long continuity. The company appears in Bangladeshi ISP-era directories from the 2000s, in old business listings, in industry member lists, in APNIC resource records, in local telecom reporting during the Ollo period, in operator-conference biographies, and in current ISPAB membership. Over time the addresses, contacts, and brand emphasis changed, but the entity did not vanish from the record. Commercially, continuity matters because it implies the company likely survived multiple telecom cycles and regulatory regimes rather than being a one-year shell.

Second, the public record proves the company controlled or was associated with scarce telecom resources: ASN allocations, wireless spectrum, and at least at one point an Ollo-linked broadband network strategy. Scarce telecom rights in Bangladesh have real option value even when a specific operating thesis fails. Spectrum, address space, gateway relationships, and current or former licenses can all be monetized, restructured, or used as bargaining chips in a concentrated regulatory market. That is one reason BIEL remains commercially interesting even though the active-exchange thesis looks weak.

Third, the public record suggests the company still touches the gateway economy. The BSCCL annual report customer line “Bangladesh Internet Exchange Ltd-IIG” is not enough to prove scale, but it is enough to reject the idea that all activity ended years ago. A company can fail to become the national exchange and still remain relevant as a customer, reseller, license-holder, or niche operator in the international-bandwidth stack. Economically, that is a very different category from a pure inactive shell.

Fourth, the record proves that Bangladesh’s exchange problem is larger than any one company. Even if BIEL had executed better, Bangladesh would still have faced the hard parts: concentrated infrastructure in Dhaka, national long-haul transport constraints, mixed incentives among small ISPs, weak route-validation enforcement, and a content-localization pattern dominated by private cache placement. The point of this is not to excuse company underperformance. It is to avoid attributing a national market outcome solely to one firm’s history. Bangladesh did not remain costly simply because BIEL failed to become the exchange. Bangladesh remained partly costly because the overall domestic interconnection market was and remains only partially cleared.

There is also a useful counterfactual here. If Bangladesh Internet Exchange Ltd today had a PeeringDB exchange object, public IX-F export, known PoPs, route-server participation, public traffic graph, visible DNS anycast or cache partnerships, and active LAN prefixes, the market would price it very differently. Investors, counterparties, and even regulators would treat it as an interconnection institution rather than an ambiguous telecom-resource company. The fact that the public record instead points to BDIX, ISPAB-NIX, and other visible NIXs is the commercial verdict already being rendered by the ecosystem.

Category recommendation

Bangladesh Internet Exchange Ltd should not be classified, on current public evidence, as an active national exchange institution of the same commercial class as BDIX or even as the newer visible NIX operators. The correct working category is narrower and less flattering: a residual Bangladeshi telecom-resource holder with historical ISP/BWA/Ollo identity, legacy ASN and address-resource traces, at least some gateway-related commercial residue, and unresolved current governance and operating visibility. It is real, it is historically important, and it may still have monetizable telecom options. But it is not publicly legible as the exchange that should have made Bangladesh cheaper.

The economics-first answer to the core question is therefore blunt. A registry-visible exchange can reduce domestic transit leakage, latency, and bargaining dependence in Bangladesh only when registry visibility evolves into market visibility. Bangladesh Internet Exchange Ltd never convincingly made that jump in public. The name captured the right economic idea, but the company’s visible history wandered into spectrum, wireless access, and resource tenure, while the actual price-discovery function settled elsewhere in the Bangladeshi peering ecosystem. Until contrary evidence appears—such as current BTRC license records, active public exchange telemetry, or renewed PeeringDB-grade operational disclosure—the rational stance is to treat BIEL as a mixed telecom-asset story, not a national interconnection platform.

The broader lesson for Bangladesh is equally clear. Making the Internet cheaper is no longer just about licensing more IXPs. Bangladesh already has eight active IXPs by public trackers. The harder problem is forcing local interconnection, domestic backhaul, content gravity, and routing trust into the same place at the same time. That requires governance visible enough for competitors to trust, security strong enough for routes to be believed, and facilities distributed enough that domestic traffic is not only “local” in law but local in path. The missing exchange in this story is less a company than a market-clearing mechanism.

Evidence ledger

BTRC NIX Licensing Guidelines https://lims.btrc.gov.bd/uploads/service_guideline/nix_guide0.pdf Source type: official regulator guideline PDF. What it supports: Bangladesh defined NIX as the domestic traffic-exchange point; required route-collector/Looking Glass capability; required public traffic reporting on the licensee website; expected consortium-style market participation and peering certification from licensed ISPs. What it does not prove: that any specific company actually complied in practice. Why it matters economically: it shows Bangladesh’s regulator understood that exchange economics depend on visible governance and traffic disclosure, not on a company name alone.

BTRC Annual Report 2018-2019 https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/9783792a025c47b9868a79aa4c4ea449.pdf Source type: official regulator annual report PDF. What it supports: BIEL received the third BWA license in 2013 with 40 MHz at 2.6 GHz; BIEL was also listed among wireless ISPs with 800 MHz and 3.5 GHz assignments; BTRC had issued seven NIX licenses by that period. What it does not prove: current operating scale, customer count, or whether BIEL’s wireless strategy succeeded commercially. Why it matters economically: it confirms BIEL became a spectrum-and-access operator, which diluted the neutrality expected of a market-making exchange.

ISPAB Public Directory Entry for Bangladesh Internet Exchange Ltd https://ispab.org/member/bangladesh-internet-exchange-ltd Source type: industry association member directory. What it supports: the company is still publicly listed as a Bangladeshi “Nationwide” member with a current contact, but with missing website, directors, PoPs, BTRC license details, and other verification fields. What it does not prove: that the company is inactive, insolvent, or unlicensed. Why it matters economically: for an exchange, missing public operational metadata is itself a negative signal because it raises operator search costs and trust costs.

APNIC Transfer Logs JSON https://ftp.apnic.net/stats/apnic/transfers/transfers_latest.json Source type: official RIR transfer log. What it supports: Bangladesh Internet Exchange Ltd was the source organization for transfers of 203.188.160.0/19 in September 2024 and 27.0.96.0/19 in October 2024, the latter to Byteplus Pte. Ltd. What it does not prove: the strategic reason for those transfers or the company’s financial condition. Why it matters economically: visible monetization or disposal of IPv4 assets is a strong clue that the company’s current value may lie more in telecom-resource management than in operating a public exchange fabric.

APNIC Transfer-Log Page Snippet on ASN131770 https://www.apnic.net/manage-ip/manage-resources/transfer-resources/transfer-logs/ Source type: official RIR transfer-log page snippet. What it supports: APNIC’s public log index shows a 2014 M&A ASN transfer of ASN131770 to Bangladesh Internet Exchange Ltd from an Indonesian source. What it does not prove: the business terms or broader transaction perimeter. Why it matters economically: it shows BIEL has historically been active in telecom-resource acquisition, not just in retail service branding.

Current ASN Mirrors for AS37994 and AS56115 https://ipinfo.io/AS37994 https://db-ip.com/as37994-bangladesh-internet-exchange-ltd https://en.ipshu.com/asn/56115 Source type: public ASN intelligence mirrors. What it supports: AS37994 and AS56115 are publicly associated with Bangladesh Internet Exchange Ltd, but appear with zero visible prefixes/IP ranges in current public mirror data. What it does not prove: the absence of all private network activity or all customer traffic. Why it matters economically: a nationally relevant exchange normally leaves richer public routing exhaust. Zero-prefix visibility supports the “thin registry trace” reading.

Historical Routing Policy Trace via Bangladesh Bank AS45532 https://bgp.tools/as/45532 Source type: public BGP/WHOIS viewer using APNIC-derived policy text. What it supports: Bangladesh Bank’s historical routing policy included import/export relations with AS37994. What it does not prove: current traffic exchange or present-day reachability. Why it matters economically: it proves BIEL once mattered enough in local routing to appear in another operator’s policy, so the company was not merely a paper artifact.

BDIX PeeringDB Record https://www.peeringdb.com/ix/2516 Source type: PeeringDB exchange record. What it supports: BDIX publicly discloses itself as open/neutral, lists facilities, LAN prefixes, traffic-stats URL, and public contacts. What it does not prove: that every listed member carries significant traffic or that all routes are settlement-free. Why it matters economically: it provides the benchmark for what a visible, price-discovering exchange in Bangladesh actually looks like.

ISPAB-NIX PeeringDB Record https://www.peeringdb.com/ix/3903 Source type: PeeringDB exchange record. What it supports: ISPAB-NIX has public contacts, a public IX-F export URL, a stated BTRC operating-license date of September 2020, and a clear neutral-exchange narrative. What it does not prove: profitability or dominance. Why it matters economically: it shows that newer Bangladeshi NIX operators understand the value of public market legibility in attracting peers.

ISOC Pulse Bangladesh IXP Tracker https://pulse.internetsociety.org/en/ixp-tracker/country/BD/ Source type: Internet Society country tracker using PeeringDB-based exchange data. What it supports: Bangladesh had eight active IXPs and 178 combined members by June 2026; IXPs exist in only one of seventeen large population centers; domestic IXP coverage reaches 59% when member/customer-cone effects are counted. What it does not prove: the exact traffic share each IXP carries. Why it matters economically: it shows that Bangladesh has exchange capacity, but not yet fully distributed market clearing.

ISOC Pulse and TBS Cache-Outage Reporting https://pulse.internetsociety.org/en/blog/2024/04/bangladesh-coping-with-submarine-cable-outage-thanks-to-indian-terrestrial-cables-local-content-caches/ https://www.tbsnews.net/tech/cache-servers-except-meta-tiktok-resume-bangladesh-903641 Source type: Internet Society analysis and local business reporting. What it supports: local caches materially cushioned Bangladesh’s cost and latency exposure; around 69–70% of top sites are locally reachable; industry sources said more than 80% of normal traffic had been served by local caches before the July 2024 disruption. What it does not prove: that IXPs themselves hosted all that cache capacity. Why it matters economically: it explains why Bangladesh’s cost problem is now partly a cache-placement problem, not only a local-peering problem.

Daily Star and Dhaka Tribune on Ollo/BIEL https://www.thedailystar.net/news/new-player-in-wimax https://www.thedailystar.net/news/ollo-to-get-lte-licence https://www.dhakatribune.com/bangladesh/bangladesh-others/56452/russian-ollo-seeks-spectrum-merger-of-its-two Source type: reputable local press reporting. What it supports: BIEL operated under the Ollo brand, pursued WiMAX/LTE/spectrum expansion, was linked to Multinet, and became embroiled in contested spectrum-merger politics with NGGL. What it does not prove: the final profitability or current ownership structure of the business. Why it matters economically: it shows BIEL’s incentives shifted into access-network and spectrum economics, away from neutral exchange governance.

Bangladesh Submarine Cables PLC Annual Report 2024 Customer List https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-bscplc/2024/12/db60f5929f4a4004b35ca564424d71e7.pdf Source type: official state bandwidth-operator annual report PDF. What it supports: a current customer line item labeled “Bangladesh Internet Exchange Ltd-IIG” appears in BSCCL’s public records. What it does not prove: full current IIG license status, service volumes, or margins. Why it matters economically: it is the best current public clue that BIEL still touches Bangladesh’s gateway economy, even if not as a visible exchange.

Watchpoints

Current BTRC license confirmation. The highest-value next check is a current BTRC license ledger for BIEL across ISP, IIG, NIX, and BWA categories. Public evidence proves historical participation in several categories, but not the present legal stack. A verified current license set would immediately reprice the company’s relevance.

RJSC and beneficial ownership. The company’s public leadership trail runs from Rafel Kabir and DNS Group-era contacts to Multinet-linked emails and Ollo-era executives, then to today’s sparse ISPAB contact. Pulling current corporate filings and beneficial owners would answer whether the company is still a telecom operating vehicle, a resource-holding shell, or part of a broader restructuring chain.

BSCCL receivables trend. One BSCCL customer-row mention is a clue; three to five years of the same line would be evidence. If “Bangladesh Internet Exchange Ltd-IIG” persists across annual reports with stable or rising balances, the company likely still has measurable bandwidth business. If it disappears, the current-service thesis weakens sharply.

Last-seen global routing activity for AS37994 and AS56115. Public mirrors say zero prefixes now, but the exact last announcement dates and historical prefix set should be pulled from route collectors and time-series tools. That would clarify whether BIEL’s routing layer died years ago or merely went private and quiet.

Infrastructure footprint outside Dhaka. Bangladesh’s IXP scarcity outside Dhaka is one of the country’s biggest structural weaknesses. Any evidence that BIEL has quietly built regional PoPs, cache partnerships, or domestic backhaul integration would materially improve its exchange thesis. Right now no public PoP list supports that.

Telecom-resource monetization pattern. The 2024 IPv4 transfers could be isolated housekeeping or the visible edge of a deeper asset-unwind strategy. Further APNIC transfers, spectrum changes, or cross-border resource moves would indicate whether BIEL’s business is increasingly balance-sheet driven rather than network-operations driven.

PeeringDB and IX-F absence. If Bangladesh Internet Exchange Ltd appears in PeeringDB or publishes IX-F-compatible data in the future, that would be a major state change. In interconnection markets, new public metadata often arrives before the traffic does. The absence of that metadata today is one reason the company should not be credited with current price discovery.