I have always followed some guiding principles when engaged in designing or reviewing customer environments. I've highlighted these before in past posts, but with last week's Citrix Summit, Mark T mentioned some of these guiding principles in his keynote and I wanted to share how they apply to the Citrix Delivery Center.
Mark's keynote highlighted the some of the issues and challenges that customers have today. For instance, today's environments are still, for the most part, in a distributed computing model. This approach makes it easy to organize people around technical disciplines, but it makes things very hard to get done in today's rapid-fire rate of change in business. I'm sure that each group is very good at what they do, but at today's rate of change it's just too difficult to keep pace with this computing model. Your agility decreases and your costs per user increases. Unfortunately, the conventional approaches most of us have used in IT in the past make it almost impossible to keep pace with the rate of change. The complexity of the modern enterprise - its business and IT components, and its linkages with other enterprises - increases the difficulty of implementing changes. Different elements change at different rates, but the pressure to change is always there. In this environment, how can we make our businesses more agile and capitalize on change?
I've always approached "change" as the fundamental architectural problem in today's environments. In discussions with clients, we recognize that to be agile we need to balance three other key dimensions: financial returns, performance, and risk management.
These high-level guiding principles promote agility and are applied across people, processes and technology throughout the extended enterprise:
Let's touch on the the first principle: Simplification. Mark had a great quote in his keynote last week from Albert Einstein; "Make everything as simple as possible, but no simpler." Very well stated Mr Einstein. Simplification is where we strive to "simplify" complex IT environments through the consolidation of applications and infrastructures, the automation and orchestration of processes, and the virtualization of resources. These tasks are at the core strategy of why you implement the Citrix Delivery Center to drive efficiencies in your enterprise.
Let's look at the tasks one-by-one:
- Consolidation of applications and infrastructure. Here is where we get away from the "deployment" scenario and get more to the "delivery" mechanism. We will centralize applications in the data center where we can more easily manage and support and control and secure applications. Mark T made the statement during his keynote, that "here is the head-end infrastructure that allows you to centralize the complexity and leave behind this very simple component where you can deliver applications". From the infrastructure perspective, I've been focusing a lot recently as this being the "virtual desktop". By centralizing the desktop experience in the datacenter we get the same benefits of manageability and support, and control and security. It also introduces other challenges as well. But in order to gain more simplicity from the infrastructure perspective this is a viable option. We can get this accomplished very well with XenDesktop. I'm in love with this product, but I want to make it perfectly clear that it is not a panacea for desktop delivery. Desktop Virtualization works best as part of a comprehensive Desktop Delivery approach which focuses on the most efficient way to get applications to the end user and meet business and IT needs, not just a PC equivalent.
- Automation and orchestration of processes. When it comes to automation and orchestration, Citrix has a good one on it's hands if it could ever get it out the door: Citrix Workflow Studio. Workflow Studio is a platform that enables simple, yet enterprise-scale, data center and business process automation through the use of Windows PowerShell. Through it's graphical workflow GUI, you can build full automation of your business and/or IT processes including user account management, server and application provisioning, security enforcement and automation, disaster recovery automation, and routine and emergency maintenance on data center resources. I've tested this product in the lab and I love it's tight integration and support for automating and provisioning applications and systems with XenServer, Provisioning Server, and EdgeSight. Now how cool would it be if visionapp vADM integrated with Citrix Workflow Studio? I spoke with Rick Dehlinger over at visionapp, since visionapp is my best practice toolset, and he tells me that depending on how Citrix finalizes the productization of Workflow Studio, visionapp plans to use it extensively. According to Rick, they will likely use it on a number of different fronts, including cross product/platform (think user/machine/workload provisioning) and user self-service/reporting. Put me on the list to help test that out Rick. I'm hopeful that Citrix has this ready for prime-time in Q4 possibly.
- Virtualization of resources. This is what everyone is talking about today: Virtualization. It's at the top of every CIOs list of priorities for 2009 and the one technology that is still seeing budget dollars being allocated to it in these tough economic times. Why virtualize your resources? According to IDC, 41 million servers will be deployed in 2010, that's a 700% increase over a 15 year timespan. The average server utilization rate still remains at 10-15%, which if you do the math, equates to about $100 -140 billion in unused capacity. Physical servers are costly to maintain. These costs encompass provisioning, housing, power, cooling, management, etc. Physical servers are static. The one workload to one server is killing companies. Believe it or not, I'm still hearing clients and potential clients saying that they are concerned with performance, complexity, the platforms supported, and (I love this one), mission critical readiness. Although virtualization's rise has been meteoric, the adoption rate is still low. We are all aware of the players here, so I'm not going to bore you with a competitive comparison. Suffice it to say, I think XenServer 5.0 has come a long way from the days of XenSource. I'll leave this up to you to decide, but everyone should be virtualizing if you aren't already.
The second principle: Standardization. In the terms of this post, I'll focus on the standardization of the Desktop and Server Operating System builds. First, let me give you a business case. I recently walked into an environment with thirty XenApp servers and every one of them was configured differently. They had a manual process for building every single server and desktop environment along with getting the right app workload on it. I'm sure I don't have to tell you that this introduces more complexity and confusion into environments. After doing an holistic assessment of their enterprise, I recommended using Citrix Provisioning Server. My roadmap for this client was to get them to a "dynamic data center". To understand Provisioning Server’s value in the datacenter, we walked through a high-level look at its key functionality – server-workload streaming. Through server-workload streaming, Provisioning Server enables that dynamic datacenter I was going for. This customer would be able to centrally manage server workloads and also deliver those workloads, on-demand. By streaming workloads on-demand rather than physically installing the workload on each individual server, their datacenters became dynamic rather than static. By doing this they accomplished the goal of "server-build consistency", which was one of their KPI's for the project. I do have one hang up here and that is the management of the standard images is not what it could be. I use my best practice tool here as well, visionapp. Their value here is they extend those built-in management capabilities.
The third principle: Integration. I actually have two definitions for this word. First, it's having an infrastructure where are the different moving pieces work seamlessly together, ie. the Citrix Delivery Center (NetScaler, WANScaler, XenApp, Provisioning Server, XenServer, etc). My second definition is to use integration to improve agility and reduce costs by dynamically linking business processes and heterogeneous, reusable IT resources both within and beyond the enterprise. Now the big important point here is the ease of configuring component services to meet new needs. For this post I'll stick with the first definition. I see the entire Citrix Delivery Center as a huge puzzle. Each and every piece fits together, where on the other hand some customers will piecemeal a solution together and introduce more complexity into their environment. By having an infrastructure that you can "modularize" only adds to the scalability and functionality that you can get out of Citrix.
These design principles (I actually left one out, but folded it into Integration, called Modularity) have established the fundamental guidelines that have shaped my customer engagements. You too can take these guidelines and use them to transform your enterprise. Having an infrastructure that is agile and flexible is built, not bought. It's a transformational process. Every enterprise arrives at the task of transforming itself with a different history, differing goals and priorities, and different transformational needs. Accordingly, for every enterprise there lies a unique path with a unique set of steps.
We live and work in a world of accelerating change. To thrive in that world we must be able to embrace change quickly, thoroughly, and efficiently. Traditional enterprise models have always emphasized just the opposite view: that change is not the norm but the exception. The reality here folks is that ever more rapid change is the norm and we must evolve our enterprise models, people, process, and technology to acknowledge the truth. In my opinion, Citrix Delivery Center can help you do this.