Summary

  • Global Telecom Brokers is best understood today as a surviving network identity inside a transferred operating business. Federal records show that Xtel acquired VDL Inc.'s communications accounts, substantially all related assets and all customer contracts in 2023; gtb.net now redirects to Xtel, which still provides GTB-specific customer resources.
  • The network identity remains visibly active. ARIN lists AS30497 and the directly allocated 69.85.0.0/18 as active, while current RIPE routing observations show 64 IPv4 /24 announcements, 16,384 routed addresses, two visible adjacent upstreams, and no announced IPv6 space.
  • Cogent and Zayo give AS30497 two visible external routing relationships, but that does not prove two physically independent paths. Both sessions could still depend on one building entrance, lateral, transport vendor, conduit, powered room or regional corridor before they separate.
  • The decisive resilience test is operational rather than promotional: who owns each access segment, who can authorize a repair, where routes and facilities share risk, how much capacity survives the largest failure, how long power lasts, which spares are local, and how quickly a qualified crew can reach the fault.

The name on the route survived the sale

Global Telecom Brokers presents an unusual but instructive picture of a regional connectivity provider after acquisition. The historic company has not simply vanished from the internet. Its name remains attached to a live autonomous system and a substantial IPv4 block. Yet the commercial operation associated with that name was transferred to another carrier more than three years ago.

The corporate and regulatory sequence is unusually clear. Global Telecom Brokers operated as the trading name of VDL Inc., a Maryland corporation. In January 2023, the Federal Communications Commission described VDL as a competitive local exchange carrier in Maryland and said it also supplied broadband internet access and interconnected VoIP to business customers in 12 states and the District of Columbia. The same FCC notice said Xtel would acquire VDL's current communications-related accounts and substantially all communications-related assets, including every customer contract.

The domestic transfer became effective on 12 February 2023, according to the Commission's grant notice. A parallel international filing said the transaction included communications assets and customer contracts, with Xtel to serve the acquired international customers under its own authorization. The FCC granted that assignment on 10 February in DA 23-132. VDL then surrendered its international Section 214 authorization effective 27 February, as recorded in DA 23-169.

The buyer's own account fits that record. Xtel's acquisition announcement described GTB as a Maryland CLEC founded in 1992, with VoIP and unified-communications customers across the Baltimore-DC-Virginia region. It said the purchase brought Xtel into the Baltimore market and added healthcare, legal, government and education customers. An adviser to the seller separately said Global Telecom Brokers sold its business to Xtel and characterised the company as a Baltimore-focused seller of voice and data services.

The public-facing migration is also visible today. Visiting gtb.net leads to Xtel, while Xtel's navigation retains resources for former GTB customers. In September 2023, Xtel said its redesigned website would give acquired GTB customers a dedicated support path. The simplest commercial conclusion is therefore not that two separate carriers now serve the same customer base. It is that Xtel acquired the operating business while parts of the technical identity remained labelled Global Telecom Brokers.

That distinction matters whenever a buyer, landlord, IT manager or regulator tries to infer responsibility from a name. The entity printed on an old circuit inventory may not be the party now answering the support call. The name returned by an IP lookup may not be the name on the current invoice. A registration can remain stable while staff, contracts, facilities and escalation rights move. None of this makes the network unreal. It makes the operating boundary the first fact that must be established before resilience can be assessed.

AS30497 is live, but a route is not a repair obligation

The strongest current evidence is at the internet-routing layer. ARIN's record for AS30497 lists the autonomous system as active under the name GTBAS. The registration dates to October 2003 and shows a last-change date in March 2012. ARIN's address record similarly lists 69.85.0.0 through 69.85.63.255 as an active direct allocation named GTBNET. That is a contiguous /18, containing 16,384 IPv4 addresses.

The registration record alone could be stale. The routing system is more persuasive. On 10 July 2026, RIPEstat's routing-status view reported AS30497 as announced, visible to all 327 of the IPv4 collectors in that observation, and originating 64 IPv4 prefixes covering 16,384 addresses. Its announced-prefix list showed each /24 from 69.85.0.0/24 through 69.85.63.0/24 continuously visible during the displayed two-week window. No IPv6 prefix was announced.

The all-/24 pattern is operationally notable. A /24 is the longest IPv4 prefix commonly propagated across the global internet, so announcing 64 of them gives an operator fine-grained control over routing policy. Different groups of prefixes can be preferred through different providers, prepended, withdrawn or shifted during maintenance. The same pattern can also be historical, or simply more specific than necessary. The CIDR Report for AS30497 observes that the address space could in principle be represented by a single /18 announcement. It cannot determine whether the 64 specifics serve a deliberate traffic-engineering need.

The crucial point is what BGP does and does not say. The protocol exchanges reachability between autonomous systems. The BGP specification lets networks announce prefixes and paths, creating a policy-aware graph of how traffic can reach a destination. It does not identify who dug the conduit, owns the pole attachment, supplies electricity to a cabinet, stocks the replacement optic or has permission to enter a locked telecom room.

A live route therefore establishes that some operating system is presenting Global Telecom Brokers address space to the internet. It does not, by itself, show which legal entity controls the routers or whose technicians will restore the local circuit. Here, a useful additional signal appears in current ARIN-derived contact data. RIPEstat's WHOIS view contains technical and DNS contacts using both legacy gtb.net addresses and xtel.net addresses. That mixed contact set is consistent with Xtel operating or administering inherited resources. It is still not a substitute for an asset schedule or a written service responsibility matrix.

This is why AS30497 should neither be dismissed as a dormant registration nor treated as proof that the pre-2023 company continues unchanged. The public record supports a narrower and more useful conclusion: the inherited routing identity remains active under a business that has moved to Xtel, and the detailed physical and contractual boundaries are not publicly mapped.

Two upstream names can hide one physical exit

Current routing observations identify two adjacent external networks. RIPEstat's AS-neighbour view lists AS174 and AS6461 on the upstream side of AS30497. Those networks are associated with Cogent Communications and Zayo Bandwidth. Independent commercial summaries, including IPinfo's AS30497 page, show the same two upstreams.

At first sight, this is the outline of a sound multihoming arrangement. If one transit provider stops carrying the routes, the other can continue to advertise reachability. If traffic engineering is well designed, the operator can distribute inbound and outbound load, perform maintenance and reduce dependency on a single commercial supplier. Two upstreams are materially better evidence than one.

But they answer a logical question, not a civil-engineering question. A BGP adjacency can be established over a remote transport service. Cogent and Zayo could be reached at separate facilities over separate laterals, which would be genuinely useful. They could also meet the network in the same room, over fibre strands in the same sheath, or through circuits bought from a common local-access carrier. Even when the two national backbones are distinct, the first ten kilometres from an Owings Mills or Baltimore customer may be identical.

There are several ways apparent diversity collapses. Two circuits can enter a building through the same underground duct. Separate fibres can share a handhole that floods or is struck during excavation. Two carrier names can lease capacity from the same metro fibre owner. A primary and backup service can terminate on the same powered aggregation switch. Distinct routes can cross the same bridge, railway corridor or pole line. Two upstream ports can sit on one edge router with one power feed. A wireless backup can share the same tower power or depend on the same congested mobile core.

The government's own resilience guidance makes this distinction explicit. NTIA's Broadband 101 material shows that duplicate fibre is not resilient when both links share demarcation points or an underground pipe; it recommends physically separated conduits, paths and building entries. CISA's resilient-power guidance similarly warns about common-mode failures between communications providers and calls for independent, geographically separated services where continuity is critical.

No public AS30497 map identifies hand-off buildings, fibre owners, ducts, metro rings or customer entrance routes. PeeringDB, a self-maintained industry database, returns no public network entry for AS30497 in its API query. That absence does not mean the network lacks interconnection. Many networks do not publish a PeeringDB profile, and private transit relationships need not appear there. It does mean that a buyer cannot use a public facility list to test whether the Cogent and Zayo paths are physically separated.

The correct interpretation is therefore conditional. AS30497 has observable upstream diversity at the routing layer. Physical diversity remains unverified. A serious resilience claim would identify the two hand-off sites, local-loop providers, entrance points, intermediate transport owners, shared-risk segments and the capacity available when either route fails. Without those facts, dual upstreams are a promising design feature rather than proof of an independent escape route.

The local bill bundles several different networks

The phrase "internet service provider" can suggest one company controlling an end-to-end system. Regional business connectivity is often assembled differently. The provider that sends the invoice can integrate access purchased from another carrier, its own metro or backbone capacity, third-party data-centre space, upstream internet transit, customer-premises hardware and a support contract. The customer experiences one service even though several organisations may own the assets.

Xtel's own contract language makes that layered structure visible. Its current terms and conditions say a data or internet service is installed when the circuit is physically installed at the customer location. They also contemplate third-party providers, special construction, extensions beyond the demarcation point and cases where facilities are not economically or technically feasible. The terms distinguish Xtel's service responsibilities from maintenance of customer or third-party phone equipment unless Xtel specifically agrees to perform the work.

Those provisions are not unusual. They are the commercial grammar of a provider that can serve many locations without owning every last-mile strand. They also show why the ownership boundary matters in an outage. Consider a fibre service to a medical office in the Baltimore region. The office may contract with Xtel. The building lateral may belong to a local access carrier. The riser may be controlled by the landlord. The edge device may be Xtel-managed. The circuit may join an inherited GTB address pool, traverse Xtel aggregation, and leave through Cogent or Zayo. Commercial power comes from the local utility, while a different contractor may be needed to splice a damaged cable.

When everything works, bundling these layers is valuable. The customer has one number to call, one invoice and one organisation expected to coordinate the chain. When something breaks, however, the difference between accountability and physical control becomes decisive. The contracting provider may open a ticket immediately but still need the access carrier to locate the fault. The access carrier may then need a utility mark-out, road permit, landlord access or specialist splicing crew. The network operations centre can reroute traffic only if another usable physical path exists.

The 2023 acquisition potentially strengthens this coordination by combining GTB's Maryland customer base with Xtel's larger support and network operation. Xtel says it owns and operates a fibre backbone monitored continuously from a network operations centre, on its partner-program page. It also advertises data-centre locations in Philadelphia, Newark and Plano, and more recent company material adds Baltimore. A 2022 Newark facility announcement describes an active-active architecture and options for diverse last-mile providers and paths.

Those statements describe Xtel's wider platform, not an as-built drawing for the inherited GTB service area. They do not show which AS30497 prefixes traverse which Xtel sites, whether a given Baltimore loop is on-net, or whether a second circuit shares the first circuit's metro route. Marketing a national service and operating a resilient local access path are different tasks. The first is a portfolio claim. The second is proved location by location.

Installed capacity is not capacity that survives a cut

Capacity language is especially easy to misread after a network acquisition. Xtel's network-solutions page advertises internet services ranging from 3 Mbps to 100 Gbps, alongside SD-WAN, Wi-Fi, 5G and cellular connectivity. Its internet-solutions page describes dedicated circuits, multiple connection types, failover through multiple data centres and service-level agreements. These are broad product capabilities. They are not evidence that every former GTB customer has a 100 Gbps port, a second circuit or multi-site failover.

The AS30497 address count is no shortcut to a capacity estimate. Sixteen thousand routed IPv4 addresses do not equal sixteen thousand subscribers, active endpoints or circuits. The block can contain infrastructure, static business assignments, dynamic pools, spare addresses and services retained for operational continuity. Likewise, 64 announced prefixes say nothing about port speed, transit commitments or busy-hour utilisation.

A useful capacity assessment needs at least three numbers. The first is installed capacity: the line rate of access ports, metro links, edge routers and upstream circuits. The second is committed or usable capacity: what the operator has actually purchased, engineered and made available after protocol overhead and traffic policy. The third is failure-state capacity: what remains after the largest credible component or route fails.

The third number is usually the least visible and the most important. A network may have two 10 Gbps upstream ports but only one 10 Gbps local transport circuit feeding both. It may have a full-size primary and a smaller backup intended only to preserve essential traffic. It may announce every /24 through both providers while preferring one path so strongly that the other is rarely exercised under load. During a failover, sessions can remain technically reachable while customers experience packet loss, latency and congestion severe enough to stop cloud applications or voice.

Xtel also makes a current claim of 99.999 percent network uptime in company announcements. Five nines would correspond to roughly 5.3 minutes of downtime in a 365-day year if measured continuously across the same defined service. But a percentage is meaningful only with a denominator and exclusions. Does it cover the core, a product family, an individual customer circuit or only on-net facilities? Are access-carrier outages, planned maintenance, customer power and force majeure excluded? Is degradation counted, or only total loss? Is the figure independently measured?

The company's legal terms are more cautious than the promotional number. They state that Xtel cannot guarantee uninterrupted, error-free internet service and that speed varies with wiring, location, configuration, congestion and remote servers. This does not negate a service-level agreement. It shows that the detailed order and SLA, not a general web claim, define what a customer can rely on. Credits after an outage may enforce accountability, but they do not replace lost clinical access, missed transactions or an unavailable call queue.

For an inherited GTB circuit, the appropriate capacity questions are concrete. What is the committed information rate at the customer port? What are the normal and peak loads on the access and upstream segments? Does the backup have equal capacity? Which applications receive priority during failover? When was a full-load failover last tested? How long did BGP converge, stateful firewalls recover and voice sessions re-establish? Without these results, installed equipment is a design ceiling, not a service guarantee.

The failure chain starts before transit

The two visible upstreams receive attention because they are measurable from outside. Most local outages begin elsewhere. A business circuit works only when every dependency between the user's device and the internet remains available. That chain includes customer equipment, internal wiring, the building demarcation, a local access medium, powered aggregation equipment, regional transport, edge routers, DNS and external transit.

The first failure point is often the premises. A fibre network terminal, router, firewall, Ethernet switch, Wi-Fi access point or VoIP phone needs electricity. Even when the provider's central equipment has generator power, a customer without a working uninterruptible power supply goes offline. Xtel's own UC legal notice warns that the product may not function during a power outage or internet interruption. That is an important reminder for any organisation that has replaced traditional lines with cloud voice.

The next point is the building entrance and local loop. Underground fibre can be severed during construction; aerial cable can be damaged by vehicles, falling trees, wind, ice or pole replacement. Telecom and cable facilities represented 47 percent of damaged facilities in the Common Ground Alliance's 2023 DIRT data sheet. That national, voluntary dataset does not measure GTB or Baltimore outages, but it shows why route maps and rapid splice response matter.

The third point is powered field and facility equipment. Ethernet switches, optical amplifiers, routers, environmental controls and access cabinets can fail even when the fibre remains intact. CISA's communications dependency primer notes that communications depend on electricity for towers, central offices and other facilities, and on transport for delivery of generator fuel. Backup power is not a binary feature. Its value depends on battery condition, load, generator start reliability, fuel endurance, refuelling access and the time needed to restore commercial power.

The fourth point is regional transport. A local loop can remain lit while a shared metro ring or inter-city path fails. A ring protects traffic only if it is closed, both directions are working, the fibre routes are physically separated and switching behaves as designed. Two labelled paths laid in the same trench are two fibres but one excavation risk.

The fifth point is the internet edge. AS30497's Cogent and Zayo adjacencies can protect against an upstream session or provider failure if the local path to each is independent and routing policy is correct. BGP itself can also fail through configuration error, route leaks, prefix filtering or router faults. The IETF's BGP operations and security guidance recommends controls including prefix filters, maximum-prefix limits, path filtering and protection of routing sessions. Public observations cannot reveal whether AS30497 implements each control.

There is also an unfilled routing-security gap. RIPEstat's RPKI validation query returns an unknown status and no validating route-origin authorization for a sampled AS30497 /24; the covering /18 returns the same result. In the terminology of RFC 6811, this is a "not found" condition rather than an invalid announcement. It does not mean the route is hijacked. It means RPKI does not provide cryptographic evidence authorizing AS30497 as the origin, so relying networks must use other policy and registration data.

The final point is human recovery. Monitoring can identify loss, and automatic routing can work around some failures, but a broken lateral still needs access, diagnosis, materials and labour. A field crew may have to test optical power, locate the cut, obtain traffic control, expose the cable, splice fibres, clean and test connectors, replace electronics and verify service. If a third-party access carrier owns the segment, Xtel's team may coordinate rather than perform that work directly. The time to repair then includes every hand-off between organisations.

Local support labour is part of the network

For a regional carrier, staffing is not an overhead separate from capacity. It is the mechanism that converts spare equipment and route design into restored service. A network with excellent monitoring but no available fibre splicer can stay dark. A warehouse full of optics in another state does not solve a same-day failure if transport is disrupted. A technician with the right test equipment cannot enter a locked rooftop or roadside cabinet without current access credentials.

Xtel publishes a comparatively specific support policy. It says the company continuously monitors its network, connects callers to a live person, routes after-hours messages to a senior technician and gives all customers same-day assistance during a major outage. These are useful commitments because they speak to response rather than only sales. Yet same-day assistance is not the same as same-day restoration, and an answered call is not a measured dispatch time.

The difference becomes larger when the fault is off-net. First-line support must determine whether the problem is customer equipment, a local access circuit, Xtel aggregation, an upstream or a remote application. If the local loop belongs to another carrier, a trouble ticket must contain enough evidence to avoid repeated tests and rejected dispatches. The owner must assign a qualified crew. If fibre is cut, the crew must find both ends and enough slack to splice. If a replacement router or optic is required, compatible hardware and the correct configuration must be available.

Acquisitions add a quieter labour risk: knowledge transfer. Long-serving GTB staff may know which customer circuits share a conduit, which legacy records are wrong, which building manager holds a key, and which route was changed during an emergency years ago. When accounts and assets transfer, that operational memory has to move into maintained diagrams, circuit inventories, escalation lists and runbooks. A live AS and functioning DNS do not prove that the full physical history was captured.

The mixed GTB and Xtel technical contacts in internet-number records suggest some continuity rather than a clean break. Xtel's continued GTB resource links also indicate an effort to preserve customer access. But the public record does not show field-team size, Baltimore-area coverage hours, splice capability, contractor agreements, spare locations or measured mean time to repair. It would be unfair to infer a shortage from silence. It would be equally unwise for a critical customer to assume that a general 24-hour support statement guarantees a qualified local crew within a particular interval.

A resilient support contract should distinguish response, dispatch and restoration. It should identify the organisation that owns each segment, the escalation clock for wholesale carriers, after-hours staffing, priority classes, spare-part policy, planned-maintenance notice and the evidence required to close an incident. For healthcare, government and education sites, it should also define how voice and emergency communications are maintained while the primary access circuit is unavailable.

Congestion can preserve the route and still break the service

Not every failure produces a clean outage. Congestion is the more ambiguous condition because monitoring may show interfaces and routes as up while applications fail. A customer can still ping an address yet be unable to sustain a video consultation, upload an imaging file, use a cloud desktop or maintain intelligible voice.

Congestion can occur at the customer port, local access segment, aggregation link, upstream transit port or remote interconnection. It can be continuous at the busy hour or appear only when a primary route fails. A backup circuit that carries no normal traffic may pass a simple availability test but collapse under the full production load. Stateful security devices can also have enough bandwidth but insufficient session capacity.

The public routing data cannot diagnose any of this. The 64 /24 announcements provide reachability information, not traffic graphs. Commercial network-measurement pages are also weak evidence. IPinfo reports a small set of responsive AS30497 addresses, inferred activity patterns and Baltimore-area locations, but those observations do not establish current customer ownership, access technology, plan speed, traffic load or test conditions. They are useful corroboration that the routes lead to responding infrastructure, not a performance record for the inherited service base.

The proper metrics are internal and time-bounded: 95th-percentile utilisation by constrained segment; packet loss, latency and jitter at peak periods; capacity after each single failure; and performance during actual failover tests. The results should be tied to the service a customer purchases. A 1 Gbps dedicated circuit should not be judged by the same contention assumptions as a shared broadband line, and an active-active design should be tested with one side removed.

This is where regional ISP economics become visible. Spare headroom, a second physical route and a larger standby circuit all cost money before they produce revenue. An operator under price pressure may rationally share transport, oversubscribe capacity or buy a smaller backup. Those choices are not automatically poor engineering. They must be disclosed well enough for customers to match price to consequence. A legal office with tolerable brief degradation and a hospital with remote clinical systems should not unknowingly buy the same failure profile.

The users at risk are business processes, not address blocks

The most credible description of the inherited customer base is institutional. The FCC described broadband and VoIP for businesses. Xtel's acquisition release named healthcare, legal, government and education organisations and placed GTB's regional activity around Baltimore, Washington and Virginia. These statements do not identify customers, and they should not be turned into claims about any particular institution. They do show the kinds of operations that the network was intended to support.

For a medical practice, connectivity can carry scheduling, cloud records, imaging, payment and voice. For a law firm, it can carry document systems, filing, secure remote access and client calls. A local government or school may depend on it for administration, public services, safety communications and teaching platforms. When voice is delivered over the same access circuit as data, one fibre cut can remove several functions at once.

The failure impact also depends on timing. A 20-minute outage at night may be absorbed. The same event during a court deadline, clinic session or emergency can be serious. A long power failure can outlast batteries at the premises or an intermediate cabinet. A regional event can prevent technicians or generator fuel from reaching sites even when spare equipment exists.

This is why outage planning should start from business processes and work backward. Which functions must continue? For how long? At what minimum bandwidth and latency? Which can move to cellular or another site? Does the backup use a different carrier, entrance and powered path? Can voice fail over to mobile devices without the office router? Are staff trained to use the alternate, and is it tested under realistic load?

CISA's communications guidance frames the same problem as interdependency. Communications support emergency services, finance, water, transport and information technology, while depending in turn on electricity, fuel transport and IT. A local business circuit is a small edge of that larger system. Its resilience is determined less by the brand on the bill than by whether the dependencies have been traced and separated.

What verifiable redundancy would look like

Global Telecom Brokers' surviving network identity does not require a public release of sensitive router configurations or exact fibre coordinates. Customers can receive meaningful assurance without exposing information that would create security risk. The useful unit is a controlled resilience statement for each service location.

First, the statement should identify the operating and ownership chain. It should name the contracting carrier, local-loop owner, backbone operator, upstream providers, customer-equipment owner and party responsible for the building riser or entrance. It should say which assets are on-net and which are wholesale. After the 2023 transfer, it should also make clear whether a circuit, IP allocation or support portal retaining GTB nomenclature is administered by Xtel.

Second, it should describe physical diversity in auditable terms. The primary and secondary circuits should use separate building entrances and demarcation points, geographically separated outside-plant routes and independent aggregation where feasible. Carrier names alone are insufficient. The provider should disclose known shared-risk groups and verify them with route letters or engineering drawings under confidentiality.

Third, it should state power endurance. That includes the customer demarcation and managed device, access or aggregation cabinets, core facilities and monitoring systems. Battery runtime should be stated at the actual load and tested periodically. Generator-backed sites should have fuel endurance, refuelling arrangements and maintenance records. A wireless backup should be tested during commercial-power loss, not only during a simulated fibre withdrawal.

Fourth, it should quantify usable failure-state capacity. The provider should show the normal committed rate, peak utilisation, standby rate and expected performance after loss of the largest path or facility. Full-load failover tests should include routing convergence, DNS, firewalls, voice registration and the applications the customer considers critical.

Fifth, it should document repair readiness. The record should define alarm-to-acknowledgement time, remote diagnosis, dispatch target, access authority, splicing capability, wholesale-carrier escalation, local spares and restoration reporting. Planned maintenance should preserve the second path rather than expose the customer to a known single point of failure.

Finally, routing hygiene should be brought up to the level implied by a long-lived autonomous system. The operator should maintain current public contacts, document intended prefixes, filter announcements and create route-origin authorizations for the AS30497 address space unless a specific technical constraint prevents it. The absence of a public RPKI authorization is not evidence of current abuse, but adding one would give other networks stronger evidence about legitimate origin.

These requirements are not extravagant. They convert broad claims such as "geo-redundant", "multiple data centres" and "24/7 support" into facts a customer can use. They also allow a regional carrier to charge appropriately for real resilience instead of competing only on headline speed.

Old listings are evidence of persistence, not present control

Global Telecom Brokers also illustrates a wider problem in evaluating small and acquired carriers: public records age at different speeds. BGP can change in seconds. A domain redirect can change in minutes. An FCC authorization can record a legal transition on a particular date. Business listings, emergency-contact databases, reverse DNS and internet-number registrations can persist for years because continuity is useful and updates are uneven.

For example, the National Emergency Number Association's company identifier search still lists Global Telecom Brokers as an active facility in Owings Mills with a 24-hour telephone contact. The Better Business Bureau has a legacy Owings Mills profile with two addresses. These records corroborate the company's historic local presence. They do not override the FCC's transaction record or establish that a separately staffed GTB facility remains in operation in 2026.

The ARIN record is more operationally relevant because the addresses are still routed, yet even there the labels do not move in unison. The autonomous-system registration retains a 2012 last-change date, while contact information associated with the organisation includes Xtel domains. The technically sensible reading is continuity through administration of inherited resources, not proof that every registry field describes the current commercial structure.

The same caution applies to location data. Commercial IP databases place observed AS30497 infrastructure around Owings Mills, Cockeysville and Baltimore. That is consistent with the company's regional history and can help detect an obviously wrong country-level attribution. IP geolocation is not a fibre map. An address can be registered at a head office, used at a remote customer site, terminated in a data centre or moved without an immediate database update. A router label or latency measurement cannot identify the conduit that reaches it.

These mismatched records are not useless noise. Together they show that the GTB identity continues to carry operational weight. The live domain transition, active routes, responsive addresses, inherited customer resources and mixed technical contacts form a coherent pattern. The mistake would be to make any one signal bear more weight than it can support.

For customers, the remedy is a dated service record rather than a search for one perfect public name. The record should pair the current contracting party with the circuit identifier, service address, local-loop carrier, demarcation, assigned addresses, support number and escalation path. It should be reviewed after an acquisition, office move, access-carrier change or network redesign. That turns a legacy label from a source of confusion into a traceable part of the service history.

A medium evidence grade, and a clear next test

Global Telecom Brokers has more current operating evidence than its sparse independent public profile initially suggests. The 2023 regulatory record establishes exactly where the business moved. The legacy domain redirects to the buyer. Xtel still exposes GTB customer resources. AS30497 remains globally visible with the full 69.85.0.0/18 divided into 64 announcements. Current registration-derived contacts include Xtel personnel, and Cogent and Zayo remain visible as adjacent external networks.

That is enough to reject the idea that the name is merely an abandoned shell. It is not enough to describe a standalone pre-sale carrier operating unchanged, or to certify a resilient Baltimore access network. The address registration is old, the physical topology is undisclosed, no public PeeringDB facility record fills the gap, IPv6 is absent from the observed originations, sampled routes have no visible RPKI authorization, and public materials do not quantify route separation, power runtime, utilisation, field staffing or restoration performance.

The fair conclusion is therefore a medium network-evidence grade. There is strong proof of historical operation, asset transfer and continuing route origination, but only partial proof of the physical system behind today's service. The single most valuable next disclosure would be a dated operating map that connects AS30497 and inherited GTB circuits to Xtel facilities, local-loop owners, independent upstream hand-offs, backup power and repair responsibility.

Until that exists, the name Global Telecom Brokers should raise a practical question rather than a nostalgic one. The question is not whether an old brand still appears in a routing database. It is whether the customer paying for local connectivity has two genuinely separate ways out, enough power and capacity to use the surviving path, and a crew with the authority and parts to repair the segment that fails.