Modern Use Cases for Application Acceleration
Riverbed recently had the opportunity to speak with a panel of industry experts—bloggers, analysts, hardcore in-the-weeds technical folks. It was an opportunity to spread the word about what we’re doing and where we’re headed with network and application visibility and performance.
We kicked off with the theme of “work from anywhere,” so it was interesting to see the Tech Field Day 22 delegates in home offices, in comfy living room chairs, or, like me, in cold, unfinished basements. We used Zoom for the event which made the theme of our presentation more palpable than even the most colorful marketing slide.
Not long ago, when we were in branch offices, we had the benefit of sophisticated network tech on the backend to make our applications perform the way they should. There was QoS on our switches and routers, MPLS with strict SLAs, high bandwidth commercial-grade internet links, direct connections to cloud providers, and WAN optimization appliances.
Rest assured, all that technology is still there. The only issue is that these days, very few people are in the office to make use of it. And, this is why Riverbed’s Application Acceleration portfolio is so relevant today.
Technology for the Way We Work Today
Application Acceleration solves the problems caused by low bandwidth broadband, DSL, satellite, LTE, and the typical connectivity we have outside the office. It improves application performance over any type of connection and for almost any application whether it’s on-premises, in the cloud, or delivered as a SaaS app.
Look at Application Acceleration as a single technology that is applied in different ways based on where resources are. Sometimes resources are in traditional private data centers, often they’re hosted in public cloud, and today many apps are delivered by SaaS providers like Microsoft, Dropbox, Slack, and Salesforce.
For years, Riverbed has made those applications perform extremely well for someone in a branch office. We used an end-to-end solution with a SteelHead appliance at the branch and another SteelHead in the data center. The results were—and still are—pretty awesome.
Branch SteelHead at the Client Level
Today, we can replace the branch SteelHead with an agent that lives right on a client computer. That means a software version of Riverbed’s SteelHead appliance is with someone no matter where they are and no matter what kind of internet connection they have.
The Client Accelerator agent is very similar to a branch SteelHead, though it’s optimized for a single computer. It’s managed by the Client Accelerator Controller—a virtual machine deployed on premises or in the cloud. This way, an IT department can manage acceleration policies all from one place.
Using the Client Accelerator Controller, we create application acceleration policies that tell the agent what to do with certain traffic. The policies look a little like firewall rules because they use source and destination IPs and TCP ports to identify traffic, though we also use URL learning and correlate local processes with network activity.
The Application Acceleration Ecosystem
Riverbed offers three Application Acceleration solutions: 1) Client Accelerator, 2) Cloud Accelerator, and 3) SaaS Accelerator.
1) Client Accelerator
With Client Accelerator, we’re not accelerating a client computer. We’re accelerating the data transfer that an application relies on. The local agent communicates with the remote SteelHead to reduce bandwidth consumption on the local link. It’ll also identify applications running on that link and apply whatever acceleration policies it receives from the controller.
2) Cloud Accelerator
In the case of public cloud, the local Client Accelerator agent communicates with a virtual SteelHead in Azure, AWS, or Oracle Cloud. A network operator can control both ends, so we still have a bookended solution that dramatically improves application performance even for cloud-hosted apps.
3) SaaS Accelerator
SaaS Accelerator leverages the same technology under the hood, but because we don’t own SaaS applications or the data centers they live in, we approach it differently. We host SaaS Accelerator in Azure and offer it as a managed service. That means Riverbed is responsible for deployment and backend management of the application acceleration service instances.
Going Under The Hood
As application traffic goes back and forth between the client and the remote server, regardless of where it is, we can pick out unnecessary packets that we don’t need to send anymore once the stream is established. We look for frequently accessed data that we can cache locally using byte-level data deduplication and data referencing. That way, we cache chunks of data and tag them with markers so they can be looked up when the client makes a request for it.
We also need to deal with the adverse effects of latency. We do that by regulating window sizing that provides a type of TCP flow control. This makes the transfer of data much more efficient. We also repackage TCP payloads to make that back-and-forth communication between a client and a server more efficient. And because we have a local agent on a client computer, we can correlate specific application processes with local network activity. Ultimately, this helps reduce round trips thereby reducing the effects of latency.
Application Acceleration is an ecosystem of components that solves the problem of poor application performance due to mediocre, sometimes outright bad quality internet connections when we’re not in the office.
In response to how we all work today, Riverbed has taken a technology we’re already experts in and brought it right down to an individual computer. And we’ve also expanded that functionality right out to the cloud—whether that’s a private cloud, public cloud, or one of today’s most popular SaaS providers.
Check out the overview of Riverbed’s Application Acceleration solution below, and make sure to watch all of our presentations from Tech Field Day 22.