To, From, and Within the Cloud—Cloud Field Day 2018 Recap
Have you ever really thought about the path a tiny piece of data takes going from your smartphone, through a bunch of infrastructure, processed by lines of code, interacted within a database, and back to your smartphone? And the fact that (hopefully) this process all happens in no more than a second? This path, good or bad, ends in what we call a Digital Experience. And, over the years, delivering a positive digital experience has become increasingly complex with the addition of cloud, containers, microservices, remote locations, and other new technologies. Just getting from start-point A to end-point B and back again, can be a journey of thousands of physical and virtual hops and miles, all depending on the architecture of your network and infrastructure and actual distance between you and “the cloud.”
Some companies provide digital experiences well, others less so. Some are able to optimize routes to, from, and within their physical, virtual, or cloud infrastructures, as well as programmatically orchestrate rules to ensure secure and high-performance network routes and transactions. Other companies rely on point solutions for network insights and management and often lack the visibility into application or network bottlenecks that are degrading the end-user experience. And who is the end-user? It can be a customer, a partner, or even an employee. Regardless, delivering superior digital experiences to any type of end user is paramount.
Understanding how the cloud works with modern networking technologies like SD-WAN, but also learning how end user experience monitoring (EUEM), Cloud Edge, application performance management (APM), and network performance management (NPM) relate to cloud infrastructure are what the delegates of Cloud Field Day #3 and Riverbed discussed. And while a small, 2-hour presentation and discussion window is barely enough time to scrape the surface of any of these topics, particularly how they specifically relate to cloud, it does lay the groundwork for future deeper dives and innovation.
Highlights from Cloud Field Day 3
Cloud Field Day consists of a series of presentations and discussions between leading technology companies and various, hand-selected subject-matter experts (“delegates”). The premise of the event is to explore a series of cloud-relevant topics, allow companies to present existing and future solutions, and engage in open-ended discussion around these technological solutions.
On April 5, 2018, the delegates met at our Sunnyvale, CA office to get a corporate overview, to understand Riverbed’s solution for managing the digital experience, to learn why data quality matters, to get insights into Performance as a Service, to see Riverbed’s SD-WAN and cloud-connectivity solution, SteelConnect, in action, and to view what’s coming next with Riverbed’s Cloud Edge Computing platform, SteelFusion.
Below are videos and highlights of each session. As there is a lot of technical information within each session, I encourage you to watch the full videos to get a complete picture.
Riverbed corporate overview with Romain Jourdan
Romain Jourdan, Director – Technical Product Evangelist at Riverbed, kicked off Riverbed’s Cloud Field Day session with a brief corporate overview, including highlighting the recent announcement that after 16 years, CEO and co-founder, Jerry Kennelly, was retiring and Paul Mountford would be the new Riverbed CEO.
Jourdan discussed the relevance of The Phoenix Project, a book on modern IT and, in essence, digital transformation. Jourdan posits that the story shows how lean thinking leads to DevOps and how companies must redefine the way they do business and how cloud now is a core part of digital transformation.
One key point of Jourdan’s is that performance is not only about speed, but also agility, how to improve user experience, how to open new locations, and business performance in general. In terms of Riverbed’s overarching strategy, Jourdan outlines Riverbed’s desire to develop, acquire, and leverage new technologies (as noted by acquisitions of Aternity, Ocedo, and Xirrus) as well as continuing to innovate around building high-performance networks to include cloud networking using a variety of technologies (SD-WAN, Cloud, WAN, Wi-Fi). Coupled to the ideas of innovations is the desire to extend digital performance to the edge while measuring digital experiences (how end users consume services, how the network behaves, how systems behave, and how the code is performing).
Managing the digital experience with Jon Hodgson
Next, Jon Hodgson, Chief Scientist at Riverbed, dove into the idea of perspectiveas it relates to performance and how based on subjective perspective, reality can be somewhat distorted depending on the role within IT: from a physical perspective the focus is on information technology while from an application standpoint, the view is more logical (e.g., services, containers, web services)—but in the end, it’s the frustration of the end user that truly impacts everything.
Hodgson outlined service delivery through a series of slides showing the flow from device to service to infrastructure and back again, illustrating how EUEM calls out problems and how having a unified NPM…
…and APM allow for much more granular and actionable insights into the full-service delivery.
Why data quality matters with Jon Hodgson
Hodgson’s next session discussed visibility challenges in the cloud. Cloud computing has introduced a “rats nest of complexities” as the technology has evolved from VMs to containers to micro-services and now to Function as a Service with elasticity and transience of these objects, said Hodgson. Also, as Hodgson pointed out, not all cloud providers expose the APIs required to accurately capture granular network and application transactions within cloud infrastructure itself. This means that visibilities and insights are only as accurate as the data that IS exposed.
In terms of understanding dependencies, Hodgson said that dependency-mapping really doesn’t work well with the complexities of modern infrastructure. As such, Riverbed decided to change the concept of dependencies to create a logical relationship and impact mapping – a performance graph. Using this methodology, one can trace and analyze from either side of the graph: from the transaction type…
…or the method.
Hodgson talked about identifying a minor performance bottleneck which, it turns out, was prevalent across the entire system. While individually this issue appeared to be insignificant (and the focus of the troubleshooting was on the slowest transactions themselves), since it existed across all topologies, it produced a compounding effect, resulting in system slowdown. Instead, using SteelCentral, ALL transactions were analyzed and the OVERALL slowest method was identified. When this small bit of code was optimized, performance across the system improved with the idea being that point solutions are unable to identify the symptom while SteelCentral’s unified solution was able to connect the dots through data interleaving and normalization and then looking at the entirety of the data.
Better performance as a service with Javier Rodriguez
Javier Rodriguez, Group Product Manager at Riverbed, next walked the delegates through the process of optimizing SaaS for performance and the process for building and configuring architectures for cloud-based infrastructures. The first principles for cloud access, as outlined by Rodriguez, are that if there is a wide-area network (WAN) between you and a SaaS server, performance can be less predictable and, in some cases, painfully slow.
To that end, Riverbed offers a managed Global SaaS Optimization which makes it easy for network professionals to accelerate SaaS application performance if and whenever it is needed. With thousands of points-of-presences distributed across the globe close to remote end users, and peer points-of-presence that are often situated less than a millisecond from core SaaS infrastructure, the Global SaaS Optimization services utilizes the industry leading SteelHead WAN Optimization technology to overcome latency, limited bandwidth or poor link quality (packet-loss) to dramatically improve the end-user experience.
In this session, Javier discusses future updates to the SaaS optimization service that will achieve even greater scale and flexibility.
According to Rodriguez, Riverbed is prototyping a SaaS performance solution with Microsoft Azure, designed for Office 365—where the Microsoft backbone is utilized and optimization is done on top of the Microsoft backbone. The solution is positioned for companies that need an easy way to boost performance and utilizes a variety of technologies including layer 7 techniques (to remove app “chattiness”), data de-duplication (elastic bandwidth) with secure support for SSL/HTTPs traffic, and network transport protocol optimizations to overcome latency or packet-loss.
Connectivity for IaaS hosted apps with Vivek Ganti
Vivek Ganti, Technical Evangelist for Riverbed, employed a mixture of hands-on SteelConnect demos with whiteboard illustrations to demonstrate how cloud connectivity can be orchestrated within and across cloud environments. With SteelConnect, it truly doesn’t matter if you have infrastructure deployed in data centers, cloud or even a mixture of clouds. SteelConnect makes it easy to create a unified “cloud of clouds” according to Ganti, allowing for even the ability to unify connectivity and management across Microsoft Azure with AWS.
Ganti showed connecting AWS Virtual Private Clouds (VPCs) to various environments, noting that the VPC structure is, in fact, not changed, but rather routes are simply added. He demonstrated how to overcome the management hurdles traditionally associated with connecting on-premises data centers to cloud infrastructure-as-a-service, even opening up the possibility of operationalizing direct-to-IaaS topologies at remote/branch sites.
Riverbed cloud edge: app deployment and protection with Parimal Puranik
Lastly in Riverbed’s Cloud Field Day presentation was Parimal Puranik, Senior Director of Product Management for Riverbed SteelFusion, who talked through the concepts and challenges involved in efficiently and securely running edge-local applications in remote offices and explained how SteelFusion provides the unique ability to centralize the management of remote data and services into data centers or the cloud, without compromising local performance or availability at remote sites.
Parimal then previewed updates to the SteelFusion cloud-to-edge storage architecture that adds support for Microsoft Azure. Plans are underway to offer AWS-based cloud-to-edge storage as well.
Digital performance for all
Cloud Field Day #3 offered a true glimpse at some ground-breaking (or cloud-forming) innovations. Riverbed presented a wide breadth and depth of networking, performance, and visibility solutions capable of spanning all types of infrastructures—physical, virtual, and cloud—as a means for optimizing business and technology performance.
We, at Riverbed, truly appreciate the time and insightful questions and discussion from the Cloud Field Day delegates and hope the talks will continue as the technology and solutions evolve.