“The real bottleneck is software. Creating software can be done only the old-fashioned way. A human -sitting quietly in a chair with a pencil, paper and laptop – is going to have to write the codes… One can mass-produce hardware and increase it’s power by piling on more and more chips, but you cannot mass-produce the brain.” – Michio Kaku.
Michio Kaku is right – hardware can be mass produced; however, just as software requires human involvement in design and creation, combining hardware to represent a functional platform does as well. After all, hardware is just components thoughtfully assembled to perform a function. We must still dictate what that function will be.
Continuing on our Beyond Citrix Experts series, in Citrix, the Hardware Layer represents the bottom rung, the platform on which Citrix and dependencies will reside. Generally, this is literally hardware – physical servers, storage, networking – with some hypervisor thrown in for good measure. Over the years, our design and deployment iterations have touched traditional multi-tier architectures where servers, networking, and storage are isolated; bare-metal and hypervisor deployments on Microsoft Hyper-V, Citrix XenServer, and VMware vSphere; hyperconverged infrastructure using Nutanix, Dell VXRail, FlexPod, Cisco UCS, and HPe HCI; and today’s hyperscale platforms, including Microsoft Azure (where, though we do not see the actual hardware, there is hardware – serverless, it is not). Networking has been just as diverse, including unique requirements for isolation due to compliance (i.e. PCIDSS), business justifications (i.e. sensitive data), and geographic distribution (i.e. overcoming latency).
In designing how Citrix relies on the hardware layer, it’s important that our architects and engineers be fluent in the requirements of diverse platforms and understand the requirements and specifications to ensure the ultimate end goal of positive user experience. Our diversity in platforms has provided us with a wealth of both experience and expertise to design and tailor solutions with and without Citrix. It also allows us to effectively speak to or against platforms based on business and technical requirements.
For example, it was a deployment of Citrix on Azure for a Fortune 100 organization that ramped our exposure to real-world, practical experience and gave way to deeply developing our expertise for deploying IaaS in Microsoft Azure. It was also our base of networking experience that effectively allowed us to deploy Citrix SDWAN to optimize both Citrix and non-Citrix traffic across continents for optimal user experience. Further, in hybrid deployments, it ensured our ability to integrate on-premises environments to Microsoft Azure, via ExpressRoute or IPSEC VPN, to ensure harmonious delivery of assets.
It was also our deep understanding of application functionality that allowed us to provide front-end optimizations, load-balancing, switching, and content control for web-based applications using Citrix ADC. As a matter of fact, you may have purchased event tickets through a platform we had built, or conducted day-to-day banking through an ADC platform we had architected and deployed.
As mentioned, hardware is rather pointless on its own – the function must still be defined where hardware becomes layered with software to perform a service, act as a platform, and provide both technical and business value. Enter the hypervisor and virtualization.
On the assumption you are aware of virtualization – VMware vSphere, Microsoft Hyper-V, Citrix XenServer – there is immense complexity in properly aligning a virtualization strategy to the hardware present or required. Industry has certainly made the design decisions and requirements around hardware and virtualization much easier, such as through the creation of validated architectures; however, unless a business has purchased design and deployment services through a vendor directly (which is common amongst hyperconverged solutions such as Nutanix and VXRail), there is still a fundamental requirement of capacity planning exercises, design and integration. The vendor will get you to a “ready” state, and the rest is on your shoulders, where you are left to the design and deployment of networking integration, resource distribution, fault tolerance, storage tiering, and related. Missteps are common at this point, and where our expertise from hardware through virtualization provides immense value, as we’re able to help properly scope, design, and deploy the solution.
At the end of the day, whether the hardware and network layer is tangible (on-premises in your data center) or intangible (accessed through a web page in Microsoft Azure), we’ve been designing and deploying hardware platforms in physical or virtual form since Day One, to not only satisfy success with Citrix but to also satisfy success for the entire infrastructure. We are experts in hardware.
Read Part III: Experts in Control