Enterprise Guide to Networking the Public Cloud – Part 1

Part 1 – Cloud Use Cases & Connectivity Options

A news release popped into my inbox last week publicizing a carrier’s new and improved solution for Amazon Web Services (AWS) connectivity.  Reading about the new offerings available for connecting to AWS peaked my interest as they are increasingly critical to our clients’ daily operations. This blog will provide an overview that is useful for both those new to the topic and people who have been involved with implementing these services since they have become more prevalent.

Below is a quick breakdown of a few basic technical use cases for public cloud services, followed by various connectivity methods made available by many of the popular cloud providers and telecom carriers. This is meant to provide the user an understanding of what use case/model might fit their existing or planned architecture and what characteristics influence the selection of connectivity methods (Cost, Flexibility, Bandwidth, QoS, Security, Complexity) that will meet their needs.

Future blog posts will go into greater detail about the various connectivity solutions including availability, accessibility, and technical designs. Check back often!

Technical Use Cases

These are the most common way to group the architectures in use when companies employ externally hosted solutions. There are countless ways in which these are put into practice in the real world, but this is meant to give a high level description along with the types of clients we see leveraging these models.

Pure Cloud – applications and data live (almost) entirely within the cloud platform

In short, this is where applications run entirely within the cloud platform and are only accessed via the Internet.  A basic example of this would be a self-contained company web site where all web code, images, databases, and/or other data reside within the cloud storage and are access solely via the Internet. We generally see this with a lot of firms that are just starting up and don’t want to invest in capital or the maintenance required to standup and host their own infrastructure. We see most companies progress to hybrid or shared clouds as they grow and the economics of hosting externally vs. internally shift.

Shared Cloud – enterprise and cloud applications live apart

In this example, enterprise applications reside within a traditional data center/colocation facility and other specific cloud-enabled applications reside within a cloud platform.  The two systems might share datasets but are not highly interdependent or interactive. Companies can leverage this model if they move a core business application to an external provider due to cost savings and new vendor offerings. Many packaged software vendors are offering hosted services where before only on-premises solutions were available.

Hybrid Cloud – enterprise applications are extended between the enterprise and cloud platforms

This scenario completely blurs the line between traditional data centers and cloud platforms.  Applications and datasets reside in either or both platforms with workloads shifting dynamically between the two platforms with client connectivity completely integrated into the enterprise network. These solutions are often reserved for mature organizations that operate some level of custom solutions and COTS applications that are integrated to enable key business processes. Most companies move to these hybrid solutions in order to change support models or to provide on-demand capacity.

Connectivity Scenarios

So how can different scenarios be leveraged to access external services either as a user or to share data between integrated platforms?  These are dictated by the applicable use case and defining the characteristics that are most important to your company (Cost, Flexibility, Bandwidth, QoS, Security, Complexity).

Open Internet – all applications accessed via Internet

Use Case: Pure Cloud

This is the most basic solution, quickest to establish and requires least amount of backend configuration.  All connectivity is performed via the Internet to publically accessible services residing on the cloud platform.  This includes publically facing services applications and backend management tools.

Internet VPN – application access split between Internet and VPN Tunnel

Use Case: Pure Cloud, Small Shared Cloud

This common solution allows for splitting of access to applications between the Internet and a VPN tunnel.  While still leveraging the Internet for connectivity, a VPN tunnel provides a second point of access for more granular security policies and improved data protected by employing encryption via an IPsec tunnel.

Carrier Virtual Port – application access provisioned via carrier provided IPVPN solution

Use Case: Medium/Large Shared Cloud

Becoming more common, this solution is provided by carriers who have existing connections to cloud providers.  They are able to provision virtual access to cloud platforms into an existing client WAN solution and provide variable levels of bandwidth based on client requirements.  This creates a well fit and tightly integrated solution where the cloud platform appears as another branch on the enterprise WAN.

Direct Peering – Application access provisioned directly between the cloud provider and client

Use Case: Hybrid and Large Shared Cloud

In a way direct peering is the most simple and flexible solution yet inversely the most complex.  This solution is provided directly from the cloud provider network where a physical interface is provisioned and connected to by the client.  Beyond this point of connectivity, it is solely up to the client to further integrate connectivity with the rest of their WAN.

Characteristic

Open Internet

Internet VPN

Carrier Virtual Port

Direct Peering

Cost

Low

Medium-Low

Medium-High

High

Flexibility

Low

Medium

High

High

Bandwidth

Low

Low

Medium

High

QoS

Low

Low

Medium

High

Security

Low

Medium

Medium-High

High

Complexity

Low

Medium

Medium

High

Augmented Connectivity – Miscellaneous and vendor dependent

There are additional forms of connectivity which represent augmentations of the solutions listed above.  One common example of this is where ISPs are publically peering directly with large public cloud providers.  No modification is required to an Open Internet or Internet VPN configuration but in theory will provide improved performance by having a more direct connection to the cloud platform.  Another example which is more popular in EU markets but gaining steam in the US is Internet Exchanges (IX) and carrier meshing services.  In this scenario a provider obtains direct peering with various cloud providers, meshes them together in a single transit area then resells client access to the meshed network transit area.  This offers a great deal of flexibility with multiple cloud provider connectivity while simplifying implementation complexity and reducing costs.

Conclusion

In this first article, we wanted to give you an overview of the current models available to customers who choose to use hosted (“cloud”) platforms for business applications and the ways in which they can appropriately connect to these services. As the number of companies, large and small, continue to leverage hosted solutions in greater numbers, the options provided through your own configuration or through a network provider will expand. It is important to know what is out there and what options make the most sense for your situation to maintain necessary performance at an optimal price.

The next part in the series will discuss the details of the Open Internet connectivity and how to leverage it in real world situations.

Your email address will not be published. Required fields are marked *

Phone: 312-602-4000
Email: marketing@westmonroepartners.com
222 W. Adams
Chicago, IL 60606
Show Buttons
Share On Facebook
Share On Twitter
Share on LinkedIn
Hide Buttons