Skip to content

Latest commit

 

History

History
341 lines (226 loc) · 21.6 KB

File metadata and controls

341 lines (226 loc) · 21.6 KB
copyright
years
2020, 2026
lastupdated 2026-04-24
keywords faqs
subcollection dl

{{site.data.keyword.attribute-definition-list}}

FAQ for {{site.data.keyword.dl_short}}

{: #faqs}

You can review answers to some frequently asked questions about {{site.data.keyword.dl_full}}. {: shortdesc}

Understanding the service

{: #faq-understanding-the-service}

How does {{site.data.keyword.dl_short}} differ from {{site.data.keyword.dlc_short}}?

{: #differentiators} {: faq} {: support}

The {{site.data.keyword.dl_short}} offering differs from {{site.data.keyword.dl_short}} on Classic in that {{site.data.keyword.dl_short}} is decoupled from classic IaaS, and exists only in the local cross-connect router (XCR). This design enables native connectivity to VPC and future capabilities without being forced into the classic IaaS network.

Direct Link allows connectivity to both classic IaaS as well as VPCs, whereas {{site.data.keyword.dlc_full_notm}} always connects to the IaaS network and a global VRF first. {{site.data.keyword.dlc_full_notm}} can only reach the VPC on a limited basis using a feature named Classic Access and by adding global routing to the direct link. See Setting up access to your Classic Infrastructure from VPC for more information.

For more information about the differences between the new {{site.data.keyword.dl_short}} offering and the classic version (Direct Link on Classic), see How do I know which {{site.data.keyword.dl_short}} solution to order?. {: tip}

Where are the {{site.data.keyword.dl_short}} offerings enabled?

{: #offerings-enabled} {: faq} {: support}

For the most up-to-date information, see {{site.data.keyword.dl_short}} Dedicated and {{site.data.keyword.dl_short}} Connect locations.

Will {{site.data.keyword.dl_short}} be available in non-MZRs, or is it only a solution for MZRs?

{: #dl-mzr} {: faq} {: support}

Initial rollout plans are for the Multi-Zone Regions (MZRs) to be prioritized. Other PoPs across the portfolio will support the new {{site.data.keyword.dl_short}} access model, enabling access to the classic infrastructure and VPC expansions as they occur.

If {{site.data.keyword.dl_short}} is not available for single-campus multizone regions, how do financial services clients handle the data that they need to keep in their regions?

{: #roadmap-szrs} {: faq}

Any existing customers on classic IaaS can remain in classic IaaS and continue to access classic IaaS data centers using {{site.data.keyword.dl_short}} or {{site.data.keyword.dlc_full_notm}}. VPC connectivity is fully supported ONLY on {{site.data.keyword.dl_short}}.

Are there documented limitations for {{site.data.keyword.dl_short}}?

{: #direct-link-vpc-limitations} {: faq} {: support}

Yes, they are listed in Known issues and limitations.

Connectivity and architecture

{: #faq-connectivity-architecture}

How does {{site.data.keyword.dl_full_notm}} work?

{: #how-does-ibm-cloud-dl-work} {: faq}

For every {{site.data.keyword.dl_short}} customer, the {{site.data.keyword.cloud}} team assigns a small private subnet to build a point-to-point network between the {{site.data.keyword.cloud_notm}} cross-connect router (XCR) and your Edge router. Then, you and {{site.data.keyword.cloud_notm}} configure the Border Gateway Protocol (BGP) to exchange routes between the environments. Finally, {{site.data.keyword.cloud_notm}} places you into a VRF to allow for the implementation of non-unique routes to the private address space of your remote network.

Is there a way to connect {{site.data.keyword.dl_short}} to VPC without using the classic infrastructure?

{: #connect-wo-using-classic} {: faq} {: support}

Yes, this is possible on {{site.data.keyword.dl_short}}. The VRF created is local to the XCR versus global on the classic infrastructure. Route targeting to VPC then enables {{site.data.keyword.dl_short}} to be used with VPC natively using the UI (without touching the classic infrastructure).

When routing on-premises subnets from the direct link through a VPC, you must create a route in the VPC routing table. For more information, see Creating a route. {: note}

How do I interconnect the classic infrastructure and VPC?

{: #interconnect} {: faq} {: support}

You can connect the classic infrastructure and VPC with classic peering as described in Setting up access to your Classic Infrastructure from VPC.

Classic access features of VPC are an option at VPC setup and can only be enabled at the initial VPC creation.

Is ECMP the best choice for redundant Direct Link connections? What alternatives exist?

{: #is-ecmp-the-way-to-go-for-redundant-direct-link-connections} {: faq}

ECMP (Equal-Cost Multi-Path) is primarily designed for load balancing across multiple links, not for providing redundancy. When using ECMP, both connections typically terminate at the same IBM Cloud cross-connect router (XCR), creating a single point of failure. Essentially, ECMP can be set up as two sessions on a single XCR.

It's important to note that you don’t have to use the same XCR for both connections. There may be scenarios involving AS Path issues similar to those mentioned in Route report considerations. Additionally, with two 10 GB direct links using ECMP, if you exceed 10 GB of throughput and one link fails, the remaining 10 GB link could become overloaded.

IBM Cloud does not recommend using ECMP in this context. ECMP load balancing only applies to traffic at the XCRs. Beyond the XCRs, the traffic from ECMP appears as the same IP address to IBM Cloud network, which defaults to the shortest path found. As a result, only one of the direct links in the ECMP configuration is actively used at a given time. {: important}

If redundancy is your goal, consider establishing two Direct Link connections——one for each XCR. For those interested in using ECMP alongside redundancy, you would need two Direct Links to each XCR to enable simultaneous ECMP sessions. Alternatively, some customers set up two links to different XCRs in the same data center, such as WDC02, and then manage failover through BGP configurations. While this approach offers some redundancy, it is less safe than having Direct Link connections in separate data centers, like WDC02 and WDC05.

Another consideration with ECMP is that if you have two VPCs advertising the same route, it might attempt to load balance across those as well. This behavior isn't limited to direct links or GREs; it can also apply to IBM Power Virtual Server workspaces. {: note}

What's the difference between the default local routing and global routing for {{site.data.keyword.dl_short}}?

{: #what-is-the-difference-between-the-default-local-routing-and-the-global-routing-add-on-for-dl} {: faq}

The local routing option is the default routing option. If your {{site.data.keyword.dl_short}} is connected at the local PoP, it provides access to all data centers within that same market. In some markets, local routing is applicable for stand-alone PoP locations and direct links that are terminated at the data center.

With our standard {{site.data.keyword.dl_short}} offering, you can send traffic between the data centers in your selected region. If you need access to other data centers outside the specified region, you must use global routing. For example, you might use global routing to share workloads between dispersed {{site.data.keyword.cloud_notm}} resources, such Dallas to Ashburn, or Dallas to Frankfurt.

Why does global routing exist for {{site.data.keyword.dl_short}}?

{: #why-does-a-global-routing-add-on-package-exist-for-dl} {: faq}

Global routing prevents you from experiencing unexpected data costs when traversing outside of your data center's local market. It lowers costs, and, if you have a global presence, allows you to reach all regions easily. However, usually you require only a local bandwidth package.

If I am connected to a {{site.data.keyword.dl_short}} in a region, such as Dallas, can I access other regions in the US through {{site.data.keyword.dl_short}}?

{: #if-i-am-connected-to-a-dl} {: faq}

Yes, you are able to gain access to areas outside of your local market if you choose global routing. If this option is not selected, your {{site.data.keyword.dl_short}} traffic is limited to the local market for the PoP or data center location you selected.

Can I connect to any available region from a given {{site.data.keyword.dl_short}} location?

{: #can-i-connect-to-any-available-region-from-a-given-dl-location} {: faq}

Yes, if you order {{site.data.keyword.dl_short}} with global routing.

Can I restrict the regions that my {{site.data.keyword.dl_short}} can reach?

{: #can-i-restrict-the-regions-that-my-dl-can-reach} {: faq}

No. {{site.data.keyword.cloud_notm}} offers two options: (1) A local market only, or (2) all regions with global routing.

How can I achieve redundancy with {{site.data.keyword.dl_short}}?

{: #how-can-i-achieve-redundancy-with-ibm-cloud-dl} {: faq}

{{site.data.keyword.dl_short}} does not provide an inherently redundant service. {{site.data.keyword.dl_short}} can provide diverse connections that enable you to create redundancy using BGP. You can achieve diversity with {{site.data.keyword.dl_short}} by connecting to more than one {{site.data.keyword.dl_full_notm}} Dedicated service provider for {{site.data.keyword.cloud_notm}}.

Technical capabilities and configuration

{: #faq-tech-capabilities-configuration}

Does IBM support IPv6 over {{site.data.keyword.dl_short}}?

{: #can-ibm-support-ipv6-over-dl} {: faq}

IBM does not support IPv6 over {{site.data.keyword.dl_short}} for the BGP session. A /30 must be assigned from IPv4, and the same is required in return from you.

Does IBM do IPV6 on the private network?

{: #can-ibm-do-ipv6-on-the-private-network} {: faq}

No. IPv6 is public only.

Does {{site.data.keyword.dl_short}} support any type of Quality of Service (QoS)?

{: #does-dl-support-any-type-of-qos} {: faq}

IBM is unable to support any QoS guarantees. QoS requires MPLS mapping between each of our service suppliers and {{site.data.keyword.cloud_notm}}. Cloud service providers generally cannot support QoS because it must reach from end-to-end and involve every device in between. No workaround is currently available by "tunneling" or any other method.

Does {{site.data.keyword.dl_short}} support Jumbo frames?

{: #does-dl-support-jumbo-frames} {: faq}

Jumbo frames (up to 9214 bytes) are supported on Direct Link Dedicated Hosting.

For Direct Link offerings, does IBM set a BGP password?

{: #on-direct-link-exchange-does-ibm-set-a-bgp-password} {: faq}

By default, BGP passwords for Direct Link aren't set up. Currently, BGP MD5 authentication is supported.

Can I request to change a VLAN ID on IBM Cloud for a provisioned Direct Link Connect gateway with an VLAN that is already allocated?

{: #change-vlan-id} {: faq} {: support}

For a direct link that was provisioned via the IBM console, a VLAN ID update is not supported. For a direct link provisioned via Provider API, you can request a VLAN update using the Provider portal/APIs, or request a VLAN update by opening an IBM Support case.

How easy is it to upgrade the bandwidth of my {{site.data.keyword.dl_short}} Dedicated connection, for example 1 - 5 GB?

{: #how-easy-is-it-to-upgrade-the-bandwidth-of-my-dl-connection} {: faq}

Typically, IBM installs speeds of 1 GB and lower on 1 GB optics. For speeds of 2-10 GB, IBM installs 10 GB optics. As a result, an upgrade of 1-5 GB would require new optics to be assigned or inserted. It would be a service affecting event. If you anticipate that type of growth, it's possible to request 10 GB optical fibers to be installed at the beginning of your {{site.data.keyword.dl_short}} deployment, or to order 2 GB initially so that the 10 GB optics are in place.

Ordering and migration

{: #faq-ordering-migration}

How do I move from {{site.data.keyword.dlc_full_notm}} to the new {{site.data.keyword.dl_short}}?

{: #move-classic-to-new} {: faq}

{{site.data.keyword.dl_short}} has a new user interface and records system, requiring you to place a brand new {{site.data.keyword.dl_short}} order.

If I order a direct link with local routing, is it possible to upgrade and switch to global routing later?

{: #order-local-routing-then-switch} {: faq}

Yes, you can change the routing option any time after creating the gateway. To do so, click Actions on the gateway's details page and then click Edit. This is not a disruptive change.

Pricing and billing

{: #faq-pricing-billing}

Where do I find cost estimates for Direct Link offerings?

{: #pricing} {: faq}

You can estimate the cost of a service using the cost estimator on the provisioning page of Direct Link offerings. For example, log in to the IBM Cloud Direct Link console and click Order Direct Link. Then, choose to order Direct Link Connect or Direct Link Dedicated. You can click the Pricing tab to get cost estimates or, as you complete the ordering form, cost estimates appear in the Summary side panel.

In terms of cost, what do I pay for?

{: #cost} {: faq}

There are two Direct Link pricing plans: metered and unmetered. Metered has a port fee and bill per GB egressed across the {{site.data.keyword.dl_short}}. Unmetered billing has a higher port fee and no usage charges, which are ideal for customers who consistently egress traffic across their direct link.

Direct Link pricing does NOT include any additional charges by service providers to enable connectivity to Direct Link.

What extra charges do I incur from other parties with {{site.data.keyword.dl_short}}?

{: #what-additional-charges-do-i-incur-from-other-parties-with-dl} {: faq}

You might have extra charges from your provider. See to your carrier or service provider for their fee information.

You, the customer, must arrange connectivity and billing with your service providers, independently of Direct Link. Direct Link creates a Letter Of Authorization / Connecting Facility Assignment (LOA/CFA) which is usable by any service provider who can reach the Meet Me Room that is specified on that LOA/CFA. The provider who is connecting to the LOA/CFA must include pricing for the cross-connect in their quote to you. Direct Link does not order cross-connects on behalf of any customer.

Can I change billing options after my direct link is provisioned?

{: #switch-billing-options} {: faq} {: support}

Yes, you can change billing options after a direct link is provisioned, regardless of whether you chose global or local routing. For example, to change from metered to unmetered billing, navigate to the Details page of the direct link and click Edit. In the side panel, select Unmetered in the Billing section, review the updated information, then agree to the prerequisites and click Submit.

When does billing begin with {{site.data.keyword.dl_short}}?

{: #when-does-billing-begin-with-dl} {: faq}

The fees for {{site.data.keyword.dl_short}} cover the cost of service termination on the {{site.data.keyword.cloud_notm}} infrastructure.

Infrastructure services are billed in advance and begin upon acceptance of a client’s order. However, due to the nature of {{site.data.keyword.dl_full_notm}}, the {{site.data.keyword.dl_short}} service billing begins when a BGP session is established with {{site.data.keyword.cloud_notm}}, or 30 days after the order is submitted.

Billing stops after (1) you request a circuit to be deleted, and (2) the provider has de-provisioned the circuit.

When are changes to my Direct Link service reflected on my invoice?

{: #billing-related-charges-on-invoice} {: faq}

Billing-related changes, such as configuration updates that affect your charges, appear on the next month’s invoice if they are made at least 7 days before the end of the current month (by 00:00 UTC). Changes made after that cutoff are reflected in the following month’s bill.

For example, for a change to appear on your October invoice, it must be made before 24 September at 00:00 UTC.

Monitoring and usage

{: #faq-monitoring-usage}

Where can you find metric or usage data?

{: #where-can-I-find-my-usage-data} {: faq}

The Direct Link offerings do not provide reporting metrics or usage data. If you need to collect metrics for a Dedicated direct link, you can collect this data from your equipment. To collect metrics from a Connect direct link, reach out to the provider for metrics if you are not able to collect data from your equipment.

What are the tools for monitoring the consumption of resources associated with the service, as well as the costs and the quality of the service?

{: #view-egress-usage} {: faq}

IBM Cloud Direct Link is integrated into the IBM Cloud usage dashboard, which provides a summary of estimated charges for all services and resources that are used per month in your organizations. This includes the number of connections and the amount of traffic flowing across your direct links. IBM Cloud Direct Link usage is billed and reported as part of the IBM Cloud invoice process.

Direct Link security with Secrets Manager

{: #enhance-security-with-secrets-manager}

What is changing in Direct Link secret management?

{: #faq-change-to-secrets-manager} {: faq}

To enhance security, IBM is integrating IBM Cloud Secrets Manager with Direct Link to manage sensitive configuration data, including MACsec keys and BGP MD5 authentication secrets. This integration ensures improved compliance and automated credential handling. It also provides Direct Link users access to a more robust platform to meet modern security and compliance standards.

Why is IBM Cloud introducing Secrets Manager for Direct Link?

{: #faq-secrets-manager-for-direct-link} {: faq}

The security landscape is evolving. Secrets Manager offers a more specialized, centralized way to manage various types of secrets, including automatic rotation, versioning, and a dedicated UI for certificate management that goes beyond standard key management. Our goal is to provide additional capabilities for Direct Link users while maintaining alignment with enterprise security standards.

What are the key benefits of moving to Secrets Manager?

{: #faq-secrets-manager-key-benefits} {: faq}

Secrets Manager provides features not fully available in standard KMS (Key Management Services), such as:

Unified Management : Manage certificates, API keys, and passwords in one place.

Lifecycle Automation : Automatically renew and rotate secrets to reduce manual overhead.

Enhanced Visibility : Detailed auditing and tracking of secret access and changes.

I already use Key Protect or Hyper Protect Crypto Services (HPCS). Do I need to switch immediately?

{: #faq-when-to-switch-over} {: faq}

No. Secrets Manager co-exists with Key Protect and HPCS until these products are announced as End of Marketing (EOM). Your current Direct Link configurations continue to function exactly as they do today. IBM is introducing this integration now so you can begin exploring its advanced features at your own pace.

What is the long-term plan for Key Protect and HPCS in Direct Link?

{: #faq-long-term-plan-for-key-protect-hpcs} {: faq}

Eventually, Secrets Manager will become the standard for Direct Link secret management. However, this is a phased transition. IBM is committed to a transparent roadmap and will provide a minimum of six months' notice before any deprecation of the Key Protect or HPCS integrations.

Does this change the price of my Direct Link connection?

{: #faq-pricing-direct-link-connections} {: faq}

No. There are no changes to the existing list prices or billing models for IBM Cloud Direct Link. This integration is a platform-level enhancement designed to improve your security experience.

How is Secrets Manager billed?

{: #faq-secrets-manager-billing} {: faq}

Secrets Manager is an existing, independent product in the IBM Cloud catalog. If you choose to use Secrets Manager, it follows its standard pricing. It is a separate line item on your bill, just as you currently see with Key Protect or HPCS.

Does this integration change how I get support or manage operations?

{: #faq-support-due-to-integration-change} {: faq}

No. Existing support channels, SLAs, and operational workflows for Direct Link remain the same. Billing for Direct Link continues through your standard IBM Cloud account. The only change is the additional option within the Direct Link UI or API to select Secrets Manager as your provider.

If I decide to migrate from Key Protect or HPCS to Secrets Manager, what are the steps?

{: #faq-migration-secrets-manager} {: faq}

See Migrating Direct Link MACsec CAKs and MD5 keys from HPCS to Secrets Manager.

What if my compliance requirements require HPCS?

{: #faq-compliance-requirements} {: faq}

IBM understands that some enterprise customers require the FIPS 140-2 Level 4 protection offered by HPCS. The Direct Link to Secrets Manager integration is being built to uphold IBM’s highest security standards. If you have specific regulatory requirements, contact your account team to discuss how Secrets Manager can support your compliance needs.