As a coffee drinker, I’m a fan of the flat white. I got introduced by a coworker a few years ago and it became my fave. Occasionally I go to my local Peets and get one. It’s a nice treat, and easy—I just click on my app then go pick it up about 5-10 minutes later.
When it comes to high-speed network services, the notion of anything close to such an on-demand consumption model was out of reach until fairly recently. With the rise of automated Network-as-a-Service (NaaS) models like what we’ve built here at PacketFabric, the days of needing to commit to inflexible, multi-year contracts to get private, optical network bandwidth services for data center, cloud, SaaS, and IX interconnection has come to an end. If you register and use our portal, you can click to deploy bandwidth on a monthly basis and get it in minutes.
But the quest to make that bandwidth ever more consumable is a journey and our mission is to keep pushing to unlock the full potential of automation so that connectivity can serve every business need. That’s why it’s exciting to showcase our hourly and burstable billing for backbone virtual circuits. You can now get high-speed, carrier-class, SLA-supported data center interconnection on an hourly basis. In fact, you can get a 10Gbps EVPL per hour for less than a flat white at my local Peets, and from the time you click to create, it’s ready to use in less time than it takes to get my coffee.
Hourly and Burstable Backbone VCs Explained
Backbone VCs are available between provisioned ports in any of PacketFabric’s ~200 global PoPs. Previously, you could choose either month or longer billing terms, or traffic usage-based billing. Now there’s a third option for hourly billing. The rest of the provisioning is the same as normal. Hourly bursting is also now an option that you can configure on top of a monthly or longer term VC. You can provision hourly-billed VCs and hourly bandwidth bursting above term-based VC via our portal or REST API. Rather than having to read how it works in text and still screenshots, you can watch this overview video.
Meeting Cyclical and Event-Based Bandwidth Demands
Hourly and burstable billing delivers bandwidth at scale with optimal price-performance for cyclical and event-based demands. By cyclical, what I mean are cyclical fluctuations in demand for bandwidth. For example, if you have an application where usage rises and falls regularly on a timed basis (daily, weekly, seasonal), you can optimize your bandwidth consumption to support application performance with a combination of term and hourly bursting. For example, you could commit your term bandwidth to cover up to a certain percentile of bandwidth requirements, then burst hourly for when demand goes over that level. With sufficient observation and API-based provisioning and deprovisioning of hourly bursts, you can tune your term baseline and bursting to deliver maximum impact and operational efficiency.
Event-based bandwidth demand is fairly straightforward. An example is migrating data from one data center to another. Depending on how much data you have to move and how timely it is, you can dial up hourly capacity up to your ports speeds (up to 100 Gbps) to get it there fast, or even choose to execute a longer-duration migration with less bandwidth with hourly or even a monthly connection.
There are many examples of bandwidth demand that occurs or changes cyclically or based on occasional events. For example, I was talking with an industry analyst recently about a healthcare organization that needed to move huge amounts of MRI machine data occasionally between data centers. Seasonality impacts IT demand in many businesses like retail, transportation, hospitality, and financial services. In Media and Entertainment, delivering production dailies for processing is another example of a cyclical requirement.
The point across all these use cases is that the choice is yours. Hourly and burstable billing offer powerful new optionals for deploying bandwidth. Speaking of cyclical and event-based bandwidth demand, there’s one use case in particular that’s an especially good fit because it is also asymmetrical.
Asymmetric Bandwidth: Backup and Disaster Recovery
Backup and disaster recovery operations require asymmetric movement of data sets. When you start, you have to seed your backup with 100% of the data set (let’s say 500 Terabytes). But that doesn’t happen again unless there’s a special event, because in subsequent backups, you’re just doing incrementals–just the 1-2% (5-10TB) of the total dataset that has actually changed. Finally, when you need to restore, you need to move the entire dataset (550TB given growth) from the backup location quickly to meet recovery time objectives (RTOs).
Backup and recovery is the perfect example of cyclical + event-based data movement. Seeding and recovery are event-driven, while incrementals are cyclical. And the extreme degree of asymmetry between those volumes of the two types of data movement have for a long time created a rather difficult price-performance tradeoff for IT teams. First of all, let’s be real that trying to seed a backup between data centers has been so challenging that in many cases the data is sent via Fedex in physical devices. So, having a bandwidth-based way of seeding is in itself a significant operational improvement.
But even beyond seeding, the problem for IT teams for networking the backup or DR operation is that you need so little bandwidth comparatively for incrementals, but the whole point of effective backup and DR is to be able to restore the lot quickly when needed. If you provision bandwidth at low levels for the 99.x% of the operational timeframe when you’re doing incremental backups, you’re going to be in serious trouble in terms of business continuity when you don’t have enough bandwidth to achieve the RTO. On the other hand, if you provision sufficient bandwidth to ensure the RTO, you’ll be paying massively more than needed for 99.x% of the time, which is hugely wasteful. Frankly, few IT teams have so much budget that they can contemplate such spendy ways.
With hourly and burstable billing, you can provision bandwidth on-demand, at the right speed needed to ensure seeding, incrementals and recovery happen in an ideal timeframe, while right-sizing your network operations costs.
Learn more or try it now
If you’re already a PacketFabric customer and have ports in two or more of our PoPs, you can provision this capability now. Watch the video above for a quick orientation, and check out our knowledge base. There’s even an hourly pricing calculator.
If you’re not yet a PacketFabric customer and you’re just starting to think through your options, request a demo or contact us at sales@packetfabric.com