Mat Mathews

Subscribe to Mat Mathews: eMailAlertsEmail Alerts
Get Mat Mathews via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Related Topics: cloudbigdataajaybudhraja

Blog Feed Post

Does HCI Really Stand for Hyper-Connected Infrastructure?

Anyone following Enterprise IT infrastructure knows of HCI as Hyperconverged Infrastructure, or the idea of collapsing compute and storage (and sometimes networking) into easier to consume and grow building blocks, for the purpose of building a cloud-like infrastructure.

Plexxi - Does HCI Really Stand for Hyper-Connected Infrastructure?http://www.plexxi.com/wp-content/uploads/2017/09/0915-hyperconnected-blo... 330w, http://www.plexxi.com/wp-content/uploads/2017/09/0915-hyperconnected-blo... 414w, http://www.plexxi.com/wp-content/uploads/2017/09/0915-hyperconnected-blo... 527w" sizes="(max-width: 300px) 100vw, 300px" />If we step back, what really is happening here is part of a larger, overall IT trend. This trend is causing the decomposition (ironically because we’re talking about convergence) of traditional, monolithic systems into easier to consume, bite size chunks. This is happening all around Enterprise IT, not just HCI. Applications are moving from large, complex monoliths, to a set of micro-services, orchestrated dynamically on demand to execute a “user experience”. Similar trends are happening with everything from data sets to large-scale data centers which are being decomposed into micro and edge data centers.

The infrastructure is actually the most obvious. Traditional SAN arrays are moving to scale-out block and file storage systems decomposed into smaller pieces that all need to continually talk to each other to synchronize meta-data and replicate data. These new decomposed systems need to work seamlessly in not just a rack or a few racks of the data center, but also across pods, data centers, and even across clouds.

As we continue to the decomposition of these fundamental building blocks, we in turn rely more heavily on the systems that interconnect them so that they can, under the command of orchestration, be used as a single entity when and where needed. This is really the age of the hyper-connected infrastructure.

What happens when the glue, the very fabric that ties together all of these disparate systems, breaks down or does not elegantly work with those very systems? It seems a bit obvious, we’re left with a lot of fragmented pieces. There remains no overall system and no dynamic infrastructure that grows and shrinks with our needs. Developers lack the speed to bring new products to market because the infrastructure needs to be thought out in advance with an idea of the worst-case scenario needed for capacity, performance, etc. in the next 3-5 years. Ultimately, businesses lose a competitive advantage over those that understand how to make infrastructure bend to their whims versus the other way around. “Let’s just move everything to AWS”, viewed in this context, is just a symptom of a disease affecting the ability of IT teams to execute the right strategic priorities. Many times, the failure to be able to leverage decomposed applications, data sets, infrastructure elements, data center locations, and even clouds, is fundamentally an inability to comprehend the overall need for connectedness.

What if the world looked different?  What if you could build an on-premise cloud starting with 3 or 4 nodes of traditional HCI (Hyperconverged infrastructure) but grow it into something as large as your data center real estate could handle if and when you needed to?  What if you didn’t need to worry about distributing your storage traffic everywhere across your network and the impact of that on your other data and application traffic? What if extending these new decomposed systems across traditional “boundaries” (racks, DCs, WANs, L2/L3, etc) was not an exercise in hand-wringing? What if the second question of every IT project was not “will our network support that”?

Turns out that connectivity is at the heart of the new IT, and nowhere is that more evident than with Hyper-converged Infrastructure. That connectivity needs to be an enabler, not a blocker. It needs to inherently and deeply understand the nature of scale-out storage systems, clustered computing systems, and multi-cloud business imperatives and just make it happen.

That is why we built Plexxi HCN™. It is, quite simply, the network for the hyper-connected infrastructure world. Plexxi HCN relies on 3 fundamental properties that makes it stand out from the old world of network imposed barriers:

  1. A world-class integration architecture, based on the open-source project known as StackStorm. From the StackStorm Git page: “StackStorm (aka “IFTTT for Ops”) is event-driven automation commonly used for auto-remediation, security responses, facilitated troubleshooting, complex deployments, and more. Includes rules engine, workflow, 1800+ integrations”. What this means is that, telling the network what you need becomes a simple act of describe what needs to happen when an event is triggered. There is nothing simpler. No amount of screen-scraping, CLI scripting, or other common hacks use to “configure” networks is simpler or more robust.
  2. An intent-based model for controlling and /computing /the underlying network, a model that we call Affinities. While every other network on the planet relies on complex, decades old protocols and modern-day “hacks” to make the network do things, ours does not. Our network was the first built from the ground up to be told what to do based on the intent of the user’s workloads. Quite simply, it is not possible to make the network behave in all the ways a business needs to accomplish its goals without having an architecture that is driven by the things that use the network. In Plexxi HCN, these things, or endpoints, are easily grouped by the user (either manually or programmatically, or magically via the integration layer), and the relationships between those groups are declared, along with their explicit needs for connectivity. Expressions such as bandwidth, latency needs, and isolation needs are done from the workloads’ perspective, not from the network device perspective (i.e. by setting meaningless QoS parameters, etc.).
  3. The world’s most dynamic physical network fabric. While most Enterprise networks force you to implement a rigid, static, physical design (such as “leaf and spine”), Plexxi HCN does not. It allows you connect your IT assets that are in the same rack, in the same pod, in the same data center, across data centers, or across clouds with equal ease and with full software control.

The hyper-connected infrastructure world is now everyone’s world. And because of that, Plexxi HCN is for everyone. Here’s an example of customers that have benefited from Plexxi HCN:

  • A government agency uses Plexxi HCN to deploy an organically grown data center environment. In their environment, new racks are deployed wherever there is space and power. They needed a network that could adapt to the physical placement of their racks, not force them to adapt to the rigid network needs. They also needed a network that was atomically driven by their VMware-based Citrix VDI infrastructure so their is no lag time between new desktop turn up and network availability. Finally, they needed a network that unleashed the power of their NetApp Clustered OnTap clustered file system, and now that they have that they see 300% faster desktop boot times, making their users very happy. That’s hyper-connected VDI.
  • Telecom Italia Sparkle, a Data Center service provider in Italy, uses Plexxi HCN to deliver cloud services to both internal and external users. They needed a network that could be managed invisibly based on their Nutanix Prism based cloud as their data centers have minimal operating staff. They also needed a network that could be expanded to their new data centers across the globe without complex network engineer skills. Click and go. That’s hyper-connected cloud.
  • Washington State University uses Plexxi HCN to allow multiple teams of data scientists to easily share data sets across the campuses multiple departments. The teams are able to keep their data in their local big data systems, but since Plexxi HCN can dynamically allocate network resources for bandwidth and latency, allow other departments to access that data as if it were local. That’s hyper-connected big data.
  • MIT Media Lab, a world class research institution and thought leader, uses Plexxi HCN to automatically users to the network without any network administrators. The network silently listens to VMware looking for new VMs to be created and automatically creates the desired network configuration and behavior.  That’s hyper-connected campus networking.

These are just a few of the examples.

Whatever your project or IT strategy is, it is likely you are dealing with the effects of decomposition of IT assets, and the resulting need for hyper-connectivity. If so, let us show you how Plexxi HCN can help you.



The post Does HCI Really Stand for Hyper-Connected Infrastructure? appeared first on Plexxi.

Read the original blog entry...

More Stories By Mat Mathews

Visionary solutions are built by visionary leaders. Plexxi co-founder and Vice President of Product Management Mat Mathews has spent 20 years in the networking industry observing, experimenting and ultimately honing his technology vision. The resulting product — a combination of traditional networking, software-defined networking and photonic switching — represents the best of Mat's career experiences. Prior to Plexxi, Mat held VP of Product Management roles at Arbor Networks and Crossbeam Systems. Mat began his career as a software engineer for Wellfleet Communications, building high speed Frame Relay Switches for the carrier market. Mat holds a Bachelors of Science in Computer Systems Engineering from the University of Massachusetts at Amherst.