If we’re honest (and we are), moving to cloud has its challenges. One of those challenges is latency. If you remember Datanauts show 16, we talked to John Merline who related this problem. Processes that used to run close by each other were now scattered across different public clouds with comparatively high latency between them. Increase network latency, increase transaction time.
Climb aboard the Datanauts train as we lay tracks to the cloud, minimizing the cars full of latency we’re hauling behind us. How will we do that? Well, this train goes to Cloud Exchange Station.
Joining us for this episode is Mark Thiele, EVP for Ecosystem Development & Evangelism at Switch SUPERNAP. Mark will discuss what a cloud exchange is, how it differs from a traditional colocation service, and how a cloud exchange addresses common challenges such as latency.
Join the Packet Pushers at Interop Las Vegas for the Future of Networking Summit, May 2 and 3. We’ll talk about the changing state of network operations, advances in network hardware and silicon, open networking, SD-WAN, containers and more. You can join us for live podcast recordings and an after-hours hangout too. Use the code PPUSHERS in the “Marketing Code” field when you register and get 25% off 5-Day, 3-Day, and 2-Day conference passes.
Part 1 – Understanding The Latency Problem
- The latency problem.
- Send. Wait. ACK. Send more.
- Bandwidth doesn’t fix latency (speed of light is only so fast).
- Where are public clouds actually located?
- How much latency are we talking about?
- Is the problem the same if I host everything with one cloud provider vs. some in one cloud and some in another?
- How much control do I have over where my workloads are geographically?
Part 2 – What Does A Cloud Exchange Do?
- What is a cloud exchange?
- Do cloud providers pay to be a part of a cloud exchange?
- How does a cloud exchange manage latency between providers? Are there special contracts in place with SLAs?
- Do cloud exchanges exist as stand alone POPs? Or are they usually part of a colocation facilities?
- Can I get a pipe from my local DC into a cloud exchange?
- Are there tiers of service available to me as a cloud exchange customer?
- Can I pay to have my traffic prioritized over another customer’s? (For example, I want my traffic to go to the public cloud FIRST once it makes it to the exchange.)
- Is this primarily about latency management, or are there other services a cloud exchange will offer me? (IP address blocks, security, reporting, SLA, etc.)
- Will I be locked into a contract?
Part 3 – Use Cases
- Who is a typical cloud exchange customer? (Enterprise, service provider, or both?)
- How much do I need my latency to drop for a cloud exchange to make sense?
- Is there any particular app or business vertical consuming cloud exchange services? Or is this service aimed at anyone with a latency problem?