Network Edge FAQ – ECX Fabric™

Topics

Key Specifications
Infrastructure and Architecture
Licensing
Connectivity Options
High Availability, Redundancy, Diversity
Regional Specifics and Roadmap
The NE Customer
Use Cases
Pricing
Quote to Order
Provisioning
Onboarding & Support

Key Specifications

What are the key components of Network Edge?

Network Edge is built on a full infrastructure stack that includes many vendors and software packages. At its heart is a compute node that we allocate CPU/processing power and memory/RAM to each virtual device, just as you would if you were ordering raw compute from an IaaS provider. Above this sits various operating systems, hypervisors, VM, orchestration services and other elements of the stack.

Equinix has heavily tested each device and pre-determined how much resources are needed for each device to operate in its optimum window. Your invoice will list a Network Edge Instance for this item.

Billable Services

Each Network Edge Instance includes several services that will appear on your invoice. Some of these services are adjustable:

Edge Instance -- the compute, memory and storage resources needed to run the devices

Virtual ECX Fabric Port – every device and Network Edge Instance comes ready to connect to the ECX platform. Since there is no need for a physical presence, the virtual port serves as the access to that service

Edge Instance Minimum Bandwidth – the Equinix Connect service provides a small allocation of public internet along with each instance of Network Edge. This bandwidth can be used to access devices and pass traffic back and forth from your private environment to the device or on to others beyond it (such as providers on the ECX). This bandwidth will be adjustable in the future

Edge Instance IP Addressing – a small quantity of IP Addresses is included to allocate to other services you may use over the lifecycle of your Network Edge Instance. This will be adjustable in the future

Other than the Network Edge Instance, you will deploy devices that are virtual images of vendor-branded services such as a router or a firewall. You can subscribe to a chosen device (Sub) to enable the license immediately or bring your own license (BYOL). If you bring your own license, there is no additional charge. If you subscribe to a license, Equinix works with the vendor to pay these costs upfront, then passes those charges on to the customer. Pricing will vary depending on which device you have deployed.

Finally, on top of the devices, select and deploy services. Most services are free, but some may incur a charge. The portal will indicate if there is an incremental charge for the service. Services include VPN Tunnels, High Availability and others. Several config customizations may be required depending on the services selected.

Does NE have a concept of a full virtual environment that I can attach many services, devices, and sub-interfaces to (such as a VNET or VPC)?

The platform is designed as a fully-functional network environment that facilitates many devices, services and use cases. The Network Edge Instance is similar to the concept of the cloud providers VNET or VPC.

How many devices can I deploy in a single location?

You can deploy as many devices as you want in a location. We will dynamically expand the resources you need to support them. from an OpenStack perspective, all of your devices are deployed in a single tenant.

A single tenant is a single deployment -- think of a single tenant like a physical colocation deployment but in a virtual cage.

What if I prefer a "non-branded" or generic solution for my use case?

The Trial Program only includes devices with a specific vendor logo and a well-known device image that can be found well-documented on their respective websites.

We anticipate adding some "generic" or Equinix-branded offers in the future. For now, the ECX Fabric features some Layer 3 capabilities that may help certain use cases. Reach out to an Equinix representative or GSA to explore options based on your individual needs.

What is the maximum allowed throughput per device that the platform allows?

Theoretically, any device on the NE platform can scale up to 10G per connection. However, many devices are closely tied to and limited by the license type associated with it. Our typical and most common licenses limit to 500M or 1G, this applies to ECX connections. We will progressively introduce other usage increments over time as the vendors and systems allow.

There are other potentially limiting factors, including Equinix Connect internet BW, the "virtual port" from NE to ECX and the destination on ECX that is desired.

How many simultaneous connections can I have on my device?

The number of connections is limited by the device type and how many interfaces it supports. A typical rule of thumb is that an average device will support up to 24 interfaces, although you should assume that a small number of those interfaces are occupied by existing services or management functions, such as SSH or the internet connection. Most users will not need that many interfaces.

Can every customer on the Network Edge platform connect with another customer?

Yes. NE offers the ability to connect to anything that the ECX can connect to; and as such, NE customers can connect to many private parties.

What SLAs are offered?

Each device is covered by an SLA of 99.99% availability. If any high availability settings are included (some of which will be available in future releases), the SLA increases to 99.999%.

One important note: if a customer makes changes to their device directly via SSH or other methods, Equinix cannot guarantee the SLA or supportability of the device once the records are not in sync.

Is A Physical Connection Needed?

No. At its core, NE is a virtual interconnection product. No physical connection is necessary.

Infrastructure and Architecture

What comprises the NE’s underlying technology stack?

Our infrastructure stack is based upon carrier grade OpenStack infrastructure, optimized for network performance.

What hypervisor is being used to deploy the virtual machines, and can I access it?

Equinix uses KVM as the hypervisor in the NE stack. Customers do not have access or interaction with this layer of the architecture.

Can I customize the software or hardware in the stack? Can I make changes to the CPU, memory, storage or other elements?

We do not allow stack customization. Additionally, the raw compute specs (CPU, memory, storage) are pre-determined per device selected. Our future roadmap may include some level of adjustment to compute within the Network Edge Instance when deploying devices if there are usage questions or customizations to the license and allowed services.

What do I need to know to make a connection to a Cloud Service Provider (CSP) on Network Edge?

You will need the same information that you’d need as an ECX customer. In general, this will involve any VLAN assignments, BGP and IP settings and unique identifiers for the specified cloud provider (Service Key, Authentication Key or account ID) as specified by that provider.

Licensing

Do I need to bring a license for the devices I will deploy in your facilities?

All devices require a license.

We offer two licensing options – Bring Your Own License (BYOL) or Subscription.

Because all devices are fully enabled branded vendor images, all customers must have a license to operate any deployed device. If you already have a license or want to purchase one from the vendor, you can Bring Your Own License (BYOL). Enter that licensing information in our portal to proceed. We also offer an ongoing Subscription immediately, so you do not need to have one in advance.

Connectivity Options

Can I connect two or more of my devices in the same metro or across metros?

At the moment there is no ability to connect two NE devices together in any scenario (sometimes called service chaining). This is a high priority item that will be coming in the future. When "remote" or across metros, it will use ECX as the reach mechanism.

Can I connect one of my own ports on ECX to a device on NEs?

The process would involve making the port on ECX a reachable destination by rolling it into a service profile. From there, you could create a connection to an ECX provider and select that profile as the destination.

Can I procure a physical cross connect from my NE-deployed device to another destination?

At the moment we cannot connect physical cross connects to your individual NE device. Although the mechanics are there to use, the method of converting that to a virtual object to pass to the virtual device does not exist. This feature is in our roadmap for future releases.

Can I connect to my device over the public internet?

Yes. A baseline amount of throughput to the Equinix Connect service is provided with every Network Edge Instance. In some cases, a customer will be able to upgrade that amount.

The typical and secure way to connect to the device is via VPN and SSH, either or both of which you can configure on your device.

Can I choose which ISP vendor I want to use to connect to my devices?

At the moment, the Equinix Connect service is the only one offered. It is automatically included with every device. In the future, we will work with other providers to include a “Bring-Your-Own-ISP” model or allow cross connects to be terminated.

How will IP Addresses be allocated? Can I bring my own IP Addresses from my network, or will Equinix license me to use IP Addresses? If I need to NAT will Equinix provide usable public address space for my use?

You can bring your own IP Addresses, or purchase through Equinix. We will provide a small number of publicly routable addresses for some features, but we encourage you to allocate pools that belong to you and that you manage.

High Availability, Redundancy, Diversity

Will there be multiple availability zones or other separation for redundant devices?

Select markets will have at least two diverse zones. We will provide an easy way to determine which markets have this available, as well as high availability choices to take advantage of them.

Will I be able to have redundant connectivity to Equinix Connect, Cloud Exchange or other interconnection products?

Yes, NE is meant to complement physical colocation solutions. Customers can always manually duplicate their setup from additional devices. Virtual deployments can always be a duplicate of a physical deployment as well, which we encourage as a use case.

Is my virtual port redundant automatically?

Yes. The administrative ports that connect between the NE platform and ECX infrastructure are in a LAG structure. Customers cannot, however, operate on or make adjustments to the structure, as this is a back-end infrastructure setup for Equinix.

Regional Specifics and Roadmap

What is the Global Availability schedule for Network Edge?

NE will roll out progressively to all regions and markets, but will not be available everywhere immediately. The current plan is subject to change, but currently includes:

Available

Silicon Valley and Washington D.C
London and Amsterdam

Planned Future Deployment

Sydney
Singapore
Frankfurt
Sao Paulo
Dallas
Chicago
Other Metros: TBD or by clear business case

Technical Capabilities & Solution

For technical questions beyond those listed below, please contact vnf-platform-support@equinix.com.

The NE Customer

The NE customer is an enterprise who requires an edge presence and desire Carrier & Cloud Neutrality. The Network Edge purchasing decision is often bundled within a larger technology project budget that’s typically driven by VPs of: Cloud Architecture, Network Infrastructure or Applications who approve the solution.

Network Engineers, IP Transport Engineers, NOC departments may all be in a position to order, configure and manage NE devices.

Use Cases

Can I use an NE device to connect remotely across markets?

Yes. Users can connect to anything that is currently reachable with the full ECX product today.

What customer screening/firewall options does NE offer for a private destination?

At the time of GA, the NE platform will feature multiple firewall brands and features that cover most all situations and needs. If you are connecting with another third party privately (versus to a public cloud destination), we always advise that you coordinate closely with that company during setup to learn their policies. Equinix will have limited visibility or control over those arrangements.

How do I connect to a partner or vendor?

Ultimately, the NE device can serve as either an A or Z end.

Please reference the ECX documentation for more detail on connecting two parties together.

Can NE help me migrate data from hardware to the cloud?

Yes. NE acts as a complement to physical deployments and the cloud. The product enables an expanding footprint for big enterprises.

Can this solution replace my physical network devices at the edge?

Yes. Deploying the virtual device does not require a physical port. NE offers the ability for an entirely virtual network if desired. We encourage use cases that deploy a virtual edge in advance of the physical deployment that may take much longer.

What is the financial break-even point between deploying my own devices in a cage (or performance hub) and virtually deploying devices on NE?

We believe this is a very important question… but also one that is radically variable depending on your business needs, priorities and circumstance. It cannot be fully answered in a short FAQ.

There are financial advantages to deploying virtual devices, but the long term outlays, projected growth, capital expenses (capex) versus operating expenses (opex) and many other factors may change that model.

Here’s one simplified example: a customer might spend $3000/month on a single Equinix Performance Hub, $1000/month for connections to various key services (such as internet access and cloud access) and $100k in capital for equipment purchases.

A sample setup might cost $6000/month for the virtual devices, licenses and interconnection services that are equivalent.

Although this may be more opex, there would be little to no capex. As mentioned, because setup and design is so radically different depending on the customer need, it is hard to determine a flat answer even to the above example. Mathematically one could run the capex payoff period and determine a satisfactory break-even. Reach out to an Equinix representative or GSA to help evaluate the pros and cons for your organization.

Who should I contact if I have questions about Network Edge use cases?

Email support is available for NE users via this address: network-edge-support@equinix.com.

Pricing

Equinix Network Edge includes the following priced elements. You’ll see these items on your invoice:

Network Edge Instance (required)–MRC and NRC dependent upon the device you are deploying and its resource requirements from small to extra-large
Edge Device License -- a Network Edge Device will always be listed on your invoice. There will only be a charge for the Edge Device License if you have a subscription license that Equinix is managing on your behalf (Sub). This is required if you do not provide your own license from the vendor.

Depending on what you'll connect the Virtual Device to, additional charges would apply from Interconnection products used.

For example, if you created a connection from your virtual device to AWS through the cloud exchange, normal pricing would apply for the ECX Fabric. The Equinix pricing guide has further details.

A full NE solution may include:

A Network Edge Instance to house devices and services
A Juniper vSRX Firewall subscription license
An ECX Connection to AWS
An ECX Connection to Microsoft

When the NE product is Generally Available, the Pricing Page on Sales Intelligence will display price lists for your region.

Network Edge Example Itemization
Invoice Item	Description
Network Edge Instance	The virtual equivalent of the cabinet you'd lease inside a Data Center. It provides the compute, storage, virtual port, internet access and IP Addressing needed to get started
Network Edge Device: License	This item displays if Equinix provides the license
Network Edge Instance - Additional Bandwidth	This item displays if you need more bandwidth than the minimum that's included in the instance
Network Edge Instance - Additional IP Allocation	We can lease additional public IP Address space for an additional fee.
ECX Virtual Connections	Pricing is per standard ECX rates and rules

Please confirm pricing either online or with your account representative.

Can I purchase devices in extended terms (e.g. 12 months) in exchange for discounted pricing?

No. At this time, the entire product is considered a pay-as-you-go pure software package. The only term is “per month”. We will be evaluating termed options soon and welcome your feedback.

Can I receive a discount when I purchase more than one device at a time?

No. We welcome your feedback on future roadmap items that would allow economies of scale in deploying several devices to a single Network Edge Instance at once.

Is there a minimum commitment to purchase?

Yes. All components of the NE platform are billed in monthly increments, so 1 month is the "minimum commit" for a customer.

What is the price to test through the Trial Program?

The Trial Program allows you to try all of the basic features and tasks that the platform has to offer at no charge. It has a fixed time limit and includes some feature limitations. Learn more about the Edge Services Trial Program.

While we do not charge to use your own license, if you opt to use your own license, your vendor may continue to charge as normal. Users will also not be charged for their connections during this period, but the size of those connections will be limited.

When I create connections in the ECX using my virtual device, who pays for the Virtual Connection?

The pricing model operates the same way as the ECX pricing model does, replacing a “port” with an NE device. Therefore, the A-end, or connection requestor pays for the connections, regardless of other factors including the destination, local or remote port or other options.

Can I apply "unlimited connections" to my virtual device to zero rate connections?

No, the unlimited connections offer only applies to physical ports.

Quote to Order

Can I deploy a device remotely (AKA in a market where I am not otherwise physically present)?

Yes. NE is designed specifically to connect to anything on ECX Fabric, regardless of the presence of physical infrastructure. We encourage this use case when deploying resources into a foreign market or one that you intend to expand into eventually.

How do I order a device on NE? How do I know how much it costs?

All aspects of the NE product are ordered online using the ECX portal or the NE API. When in the portal, the system will indicate the price you will pay as you make changes to your solution before placing the order.

Can I reserve infrastructure for my devices in advance of deploying the device?

At this time, you cannot order a Network Edge Instance before requesting a device. This feature may be available in the future. Once you request a device, we allocate the infrastructure required to run your device with a Network Edge Instance.

The Network Edge Instance space that is allocated to "hold" the device is dedicated to you. At that point, those resources are removed from the available pool for all other customers.

If my devices are using most or all of the Edge Instance resources, can I allocate more resources to them?

At this time, you cannot make changes to the building blocks that make up a Network Edge Instance, but this feature may be available in the future.

Provisioning

For onboarding & support questions that aren’t addressed in this document, please reach out to an Equinix representative or vnf-platform-support@equinix.com.

Can I log into my device and configure it manually or in a traditional manner from command prompts with TACACS or other means?

Yes. All devices have the option to load an SSH service that includes setting up user credentials and specific IP Addresses the customer would like to connect from. Using industry standard tools like telnet or PuTTY, customers can access the device and perform nearly any command that is possible from the vendor’s user documentation.

This includes many settings that will not be available directly from the portal or API as a service overlay or automation. Customers who do this should be aware that once changes are made with SSH, Equinix support cannot guarantee that functions will work as expected, synchronization will continue, or that they will be able to troubleshoot and isolate issues.

Can I load my own preset configuration templates onto my device that match my companies approved standards?

At this time you cannot load your own device images or a unique version of the OS for your device vendor or manufacturer. This feature is in our roadmap for future releases.

Can I configure and deploy my device with a tool other than the NE portal?

Some devices and vendors allow us to use their own software. We enable that -- you can connect to it and disable services on our portal as a result. To determine if that configuration is an option for you, please coordinate with your vendor.

Can I clone a device already deployed elsewhere on NE?

You cannot save a device image and then use it as a backup, copy or reload. Multiple ways to accomplish this are on our roadmap for future releases.

Trial Program Support

Email support is available for NE users via this address: network-edge-support@equinix.com.

How are vendors and their devices approved for use?

Equinix conducts a thorough onboarding process for each vendor that includes product, network architecture, engineering and service orchestration. The vendor’s device must comply with a variety of minimum viable standards before integration into our platform.

We also conduct a battery of testing scenarios to check for successful operation, optimal resource allocation to the virtual device, ability to support and determine settings and performance, interoperation with our orchestration and many other tests before a virtual device is added to the platform for customer use.

Can Equinix manage the devices I deploy into NE?

Yes. Like any other network service, Equinix Operations personnel are empowered to see certain information about your device and assist with troubleshooting.

Our systems continually monitor your device for certain statuses and conditions at all layers of the technology stack. We prefer to have explicit permission from the customer before actually logging into the device itself, just as we would a physical device in a customer cage.

What should I do if customers have questions on Network Edge’s capabilities or suitability to their needs?

Customers can reach out directly to their cloud engineer with any questions on device capabilities or further connectivity requirements.