Planning and Developing an IT Hardware Strategy utilizing virtualized hardware.

MIS535 Management Applications of Information Technology

Table of Contents

Abstract 3
Brief Company Background 4
Discussion of Business Problem 5
High Level Solution 6
Benefits 8
Conclusions and Overall Recommendations 11
Summary 13

Abstract
Corporate users of technology use a variety of technologies to get their work done. Over the last three decades, companies that create and manage data have implemented a wide variety of disparate computers systems, each slightly different. This makes maintaining these environments expensive and troublesome to expand and maintain. A potential solution to this multi-platform server sprawl is to consolidate from many to a few computing platforms. Once the application software is ported to these new platforms, we can begin to virtualize hardware. These are the prerequisites for cloud computing. This plan addresses these concerns in my current place of employment which has been experiencing this problem for over a decade. Based on industry publications and communications with my peers, it appears that many companies are experiencing these types of issues. The recession of 2007-2008 has been very difficult for IT firms. We’re doing more with less and it appears that the budgets won’t be significantly increasing for some time.

Company Background
The name of the firm in this project is the Depository Bank and Trust Corporation of New York. This firm is an FDIC member and a clearing house for Wall Street firms. The bank is in the process of expanding internationally and has gone from a single US datacenter, to maintaining offices in the Netherlands, Singapore, Shanghai, Chennai and Tokyo. Rapid expansion has created a fragmented Information Systems infrastructure, which has made the environment difficult to manage. Costs have also escalated to the point where they need to be contained. In today’s extremely competitive environment, we need to plan and execute cost effective, reliable solutions to support our international expansions. We need to use local partners to host our datacenters and conform to very stringent multi-country regulations. We need to be the best cost, most reliable solutions provider to the industry.

Business Problem Statement
In the 1970’s, there was no computer interoperability, you went to your vendor, typically a monolithic computer vendor like IBM and you purchased a turnkey solution to your computing needs. You may have needed to change your business processes to adopt to your new computing environment. The overhead required to computerize a medium to large sized firm was staggering. You were partnered with your vendor, with little way out of relationship. In the 1980’s competition crept up on the large vendors. The advent of the personal computer changed everything. Companies who had large investments in hardware and software were now in competition with personal computers running Lotus 123. These isolated computers posed no threat to the large computing environments until mid-decade when the Local Area Network became available for small computer system. Two decades ago, as computers became more connected, interoperation between vendors was the driving factor for hardware vendors. Standardization of protocols like TCP/IP made it possible for servers from different vendors to communicate. Companies like Sun Microsystems and IBM were big winners with their Solaris and AIX operating systems. Santa Cruz Operations (SCO) also had a UNIX flavor that ran on X86 based hardware. A few years later, companies like Microsoft and Novell made great strides in both the desktop and enterprise environments. When we fast forward into the present, we find that we have a majority of legacy systems, running on various proprietary platforms. We have decided to reduce our Operating System palate to just two. Windows and Linux! Windows2008 Enterprise R2 or Red Hat Enterprise Linux 6 will be the only two choices. This greatly reduces our hardware and software support costs and allows virtualization of all future servers, creating many future synergies. Today in 2013, many of the lower tiered servers have already been consolidated and/or virtualized. However, virtualizing business-critical applications—databases, ERP systems, email servers, and industry-specific solutions—feels like a completely different ballgame, and there are questions whether the risk vs. return ratio is still worth it. In addition, these applications often have application owners and database administrators that may not be familiar with virtualization. (VMware, 2012). These issues need to be addresses before we can proceed with the implementation.

Proposed Solution
We will first need to get an accurate inventory of all of the company server assets. Once we understand our inventory, we can see what needs attending to. We need to communicate what virtualization means to all of the stakeholders at the firm. As Kenneth Lauden, (2012) mentions, Virtualization is the process of presenting a set of computing resources so that they all can be accessed in ways that are not restricted by physical configuration or geographic location. Virtualization enables a single physical resource such as a server or storage array, to appear to the end user as multiple logical resources. David Marshall (2011) mentions that stakeholder buy-in should be straightforward once the top benefits of virtualization are communicated. Among the top benefits are faster server provisioning, which reduces our time to market and gives us a strategic advantage. We’ll all take advantage of better disaster recovery scenarios with virtualization. Presently, each application has an individual disaster recovery scenario. With many different systems, dependencies and different priorities make DR a difficult, if not impossible task. Virtualization solves this. Virtualization also gives our stakeholders an added benefit. They can, if they choose to, extend the life of older applications. Some parts of our organization may be locked in to an old solution currently running on obsolete hardware. Without a budget to replace these legacy apps, virtualization may be the only option to keep them alive. Once the hardware dependency is removed, the virtualized application may run on modern, backed up systems indefinitely. Once virtualized, these apps also have the ability to move into the cloud, which gives the firm a new set of capabilities. These applications can be in the cloud, yet isolated from other systems. Once the virtualization effort is completed, uptime for these applications will be increased. As an example, our mail center uses an old Palm based application for internal mail deliveries. This PC based system is over a decade old and the vendor has been absorbed into a larger firm. Once these servers are virtualized, the life of the system is extended to the life of the Palm pilot devices, of which we own over five hundred. Some of the other benefits are of interest more to IT infrastructure than the stakeholders. Among the benefits , a large reduction in the datacenter footprints. With many new international data centers coming on-line, the firm will be able to reduce the company’s energy footprint, This means far fewer physical servers, less networking gear, a smaller number of racks needed. In addition, data center cooling and facilities management is greatly reduced. IT will also benefit from server consolidation and virtualization with easier QA and development environments, each on its own isolated network. These efforts also reduce vendor lock-in. As mentioned previously in the abstract, in the early days, companies were completely locked-in to vendor’s specific solution. Being tied down to a particular vendor can be quite frustrating, since negotiating power may not be effective. Since server virtualization abstracts away the underlying hardware (Marshall 2011) and replaces it with virtual hardware, data center managers and owners gain a lot more flexibility when it comes to the server equipment they can choose from. This can also be a handy negotiating tool with the hardware vendors when the time comes to renew or purchase more equipment. These powerful benefits make it easier for stakeholder adoption. Once the buy-in in complete, we’ll need to select the vendors and partners that will provide our virtual infrastructure. We’ll need to make it highly available, spread it out internationally, and be able monitor and operate it in an efficient manner. Once we have a list of candidates we’ll prioritize which virtualization candidates will be transitioned to virtualization. Application development will need to make code changes to software to make it run on RHEL or Windows 2008. Since we’re already running java for most of our infrastructure, the transition shouldn’t be too painful. We’ll need to closely manage the virtual environment for capacity, and create a shadow QA environment for software and hardware testing.

General Benefits
We currently have multiple hardware vendors’ servers, running a large variety of Operating Systems, with application code written in languages from assembler to COBOL. Standardization to RHEL and Windows2008 with C++ and Java will give us the ability write portable code that will run anywhere in the infrastructure. If we look at server utilization in the datacenter, we see low utilization across many servers. With Virtualization, we’ll have fewer servers, better utilized, with a smaller datacenter footprint and higher availability and supportability. We’ll also be faster to market and much more dependable Data Security issues will be improved in our new virtualized environment. All of the top benefits mentioned in the solution will be leveraged upon. There are also a number of other concerns that need to be addressed, security being the most urgent. A critical requirement of datacenters in todays cloud-based computing and services environment is that they provide a common infrastructure serving multiple groups or customers (Microsoft 2012) all the while keeping each group’s data private and secure by enforcing full isolation of each workload from all other groups’ workloads. Multitenancy, as it is known, provided a good level of workload isolation between virtual machines in server virtualization, but recently the network layer of the virtualized datacenter was still not fully isolated. The need for multitenant security and isolation is key. This provides the flexibility to restrict any customer’s access to a virtual machine on any node while still maintaining network and storage traffic isolation. The firm needs the controls to isolate important “keep the lights on” mission critical traffic from other day-to-day network traffic. Access controls must also be comprehensive and manageable. One of the other new concerns in this virtual, cloud ready environment is server sprawl. Since the servers are easy, fast and cheap to provision, a large quantity will be ordered. As Techtarget.com (2009) mentions, virtual servers have the same needs as physical servers when it comes to security, storage and management. If you can't control VM sprawl, there can be serious ramifications on the safety and reliability of not just your virtual server infrastructure, but of your physical infrastructure too. When the number of virtual machines (VMs) on the network reaches the point where an administrator can't manage them effectively -- or where the VMs start demanding excessive host resources, that means there is virtual sprawl. There needs to be a formal request system, based on our architecture office, to determine if the servers requested are actually necessary, or if their functions can be consolidated into existing resources. Although this will slow down provisioning speed, it is critical if we want to plan for growth. As companies have adopted virtualization, it has become easier and easier to create virtual servers. What used to take weeks can be done in minutes. But now the bill is coming due and companies are dealing with virtual server sprawl. Virtual sprawl is really a class of problems related to the uncontrolled creation and ongoing lifecycle of virtual machines. Virtualization by its nature is extremely dynamic and flexible, and with anything that is extremely dynamic, and even disruptive, there are always new concerns and new considerations.

With virtualization you can create, copy, change and move things very quickly. But what has not kept pace is the ability to track and manage these virtual assets. There are massive blind spots in the virtual world, areas where virtual assets are used and can grow, and then fall into disuse without being monitored by virtualization technologies like VMware, or by traditional data center management systems. Everybody is overspending as a result of all of these new dynamics.
Many customers complain that virtualization does not save money, but actually costs more that keeping physical servers. Forbes Magazine’s Dan Woods (2009) did an expose on what he found. The most critical take away to virtualizing our firm is again mentioned by Forbes, Most overspending is related to the hard cost of storage for machines that aren’t being used at all. The number is actually much larger when software licensing, support costs and fractions of head count are factored in. In fact, the majority of adopters overspend on server sprawl. The company will need to take a close look at controls and processes in the datacenter. It will need to use COBIT, ITIL or whatever processes is in place to insure that best practices for virtualization and cloud computing are in place. Growth and capacity are to be closely monitored to insure that the computing resources are healthy and will have a long, productive lifecycle.

Conclusions and Overall Recommendations.
Our firm has been thrown into an environment where globalization is our only method for growth. In order to survive, we need to quickly abandon our old business center based IT provisioning to a unified manageable, scalable solution. We need to abandon legacy platforms and focus on virtualization on industry standard platforms with an eye on cloud computing. Our infrastructure has to be able to support 24/7 operations on a global scale with no down time.
We have standardized our server platforms to Red Hat Enterprise Linux (RHEL) and Windows 2012. These platforms have the largest ecosystems around them and application developers are plentiful. We have also standardized on Hewlett Packard Proliant Hardware as our Server vendor. HP has a huge global footprint and an entire support division for mission critical systems. For our Networking hardware, Cisco System’s Nexxus line of switches and routers give us world class speed and reliability. Like HP, Cisco has a huge global footprint and can support us 24/7. For our storage EMC gives us many storage options for the new environment. Everything from NFS based storage for our non-critical systems to the fastest Fiber-channel devices for our most critical data. We chose VMware for our virtual server Hypervisor. In addition to being the reliable and robust system, VMware’s vSphere offering gives us the tools needed to manage our environment.
With these vendors in place, we’ll need to change our internal processes to best leverage out new environment. We’ll need to adopt new best practices to insure we don’t fall prey to any of the major problems that have manifested themselves in the industry. Well firstly need to take a cold hard look at security, and come up with plans that meet our needs. We’ll need to develop best practices around access controls, along with isolating data and traffic based on importance. Secondly, we’ll need to do all we can to minimize server sprawl. Again business processes need to be changed and best practices will need to be implemented to insure we don’t overspend on needless resources that will raise costs. We’ll need to be “stingy” with the average server. Our internal customers will need to justify larger, or more numerous servers with processes that prioritize who gets additional resources. Thirdly, we’ll need to control and closely manage growth of our new global IT infrastructure. We’ll need to install tools to monitor capacity and faults. VMware provides a package called Capacity IQ which monitors usage trends and anomalies in the compute infrastructure. EMC also provides tools for storage use. We’ll need to insure that we are involved in the budget process and we will need to forecast what our usage is one to three years out. We’ll need to keep abreast of trends both in computing and banking so that we’re not blindsided by an industry event. Fourthly, we’ll need to protect our new infrastructure from both hardware failure and disaster. As the events of 9-11 have shown our firm, you cannot be prepared for everything. Before 9-11, we had two datacenter in downtown Manhattan. Both were on the same power grid as the World Trade Center. It took three days to install emergency power and bring up a minimal group of computers. In our new global, infrastructure, we should be able to move “applications” anywhere we have a presence. We can bypass disasters as they happen with minimal impact to our firm and customers.

Summary

Reduce number of computing platforms to two.
Develop roll-out plan.
Select Vendors. Partner with best-of-breed.
Get stakeholder buy in.
Virtualize “low hanging fruit” servers.
Monitor results and change accordingly.
Develop Enterprise Infrastructure best practices and procedures.
Virtualize “tier 1” servers.
Minimize Server Sprawl.
Test Backup.
Test DR.
Monitor and address faults.
Monitor Capacity.
Forecast growth intelligently.
Ensure proper budget to enable growth.
Keep a close eye on security.
Enjoy the benefits of a secure, stable, global infrastructure.

References

Laudon, K.C., Laudon, J.P. (2012) Management Information Systems. Custom Edition for Devry University. IT infrastructure and emerging technologies P 182

Marshall, David, (2011) Infowold, Top 10 benefits of server virtualization http://www.infoworld.com/d/virtualization/top-10-benefits-server-virtualization-177828 Microsoft Corporation (2012) Server Virtualization with Windows Server 2012 Hyper-V http://download.microsoft.com/download Tech Target (2009) Controlling VM Sprawl in your infrastructure http://searchservervirtualization.techtarget.com/Control-VM-sprawl-in-your-virtual-server-infrastructure VMware, (2012) Virtualizing Business-Critical Applications on vSphere White paper 2012 http://www.vmware.com/files/pdf/solutions/VMware-Virtualizing-Business-Critical-Apps-on-VMware_en-wp.pdf Woods, Dan (2009) The costs of Sever Sprawl, Forbes Magazine, Web edition
http://www.forbes.com/2009/12/01/embotics-computers-enterprise-technology-cio-network-virtualization.html