Data Center Capacity Planning

Chris Haynes[Senior Systems Consultant]

Capacity PlanningIs your data center infrastructure being used the way you intended it to? Are you in control of your organization’s IT data?  If your organization is like most, probably not as well as you would like to.

A good data center capacity planning strategy can help your business become more efficient, dynamic, and competitive, but few organizations dedicate resources to capacity management or give it any importance.  If it does exist, many times it is usually just to satisfy regulations or to check a box on a form.

The tendency of many IT departments is to play it safe by purchasing more server and storage capacity than needed, which means resources are being wasted at a time when IT departments are being asked to do even more with less.  On the flip side of the coin, some IT departments are not able to anticipate the growing IT demands of the business, and react to them in a timely manner, potentially resulting in lost business or costly reactionary purchases.

In order to proactively meet the demands of your organization’s needs, you need to develop data center capacity planning processes and procedures that will track and report on the following metrics:Capacity Planning 2

  • How much is being used?
  • Who is using it?
  • At what rate is it being used?
  • How much is left?
  • When will we run out of it?

Having the answers to these questions at any point in time allows your IT department to be proactive and take steps accordingly to stay ahead of the business’ needs.  Being able to project future needs allows your organization to budget and plan for data center growth and scalability in advance, which is almost always cheaper than unplanned emergency purchases.

Does your organization need help developing a data center capacity management strategy?  Please contact your Keller Schroeder Account Manager to see how we can help.


Put Your Data Center on a Diet

Kris Linville, Sr. Systems Consultant


Bikini season is just around the corner. Is your data center ready for its pool-side reveal? If you’ve spent years packing your data center full of sweet devices and high-calorie cables that have your racks bulging at the waist line, it’s time to put your data center on a diet…the UCS diet.


Data centers grow quickly in number of servers due to departmental application requirements. Maybe the HR department or the Accounting department is gearing up for a brand new system that would require multiple application and database servers for load balancing and redundancy. Is there enough rack space to install all of that new hardware? Is there space to add additional racks if needed? Obviously, it doesn’t take long for a data center to start feeling the effects of packing on those extra pounds time after time.


UCS-racks-beforeVirtualization can be leveraged to quickly get you back down to your target size. Reducing many physical servers to just a few VMware hosts is a great way to kick-start your diet. However, over time you will notice that as your admins and users get comfortable with their new diet, combined with how easy it is to deploy more virtual servers, your waist line will slowly start to creep up again. More hosts will be required to support your virtual workloads and before you know it your once-empty racks will be full again. How can you mitigate the effects of this “yo-yo” diet?

UCS-racks-after

Cisco Unified Computing System (UCS) is the solution to your virtualization “yo-yo” diet. Combine the space and administration efficiency of virtualization with the resource density offered by Cisco UCS B-Series Blade Servers and you will finally be able to maintain the ideal trimness you’ve been longing for. UCS is designed to create high-density, scalable environments that can reduce an entire data center of physical servers into a single rack.



By condensing up to 8 physical rack servers into a single 6U UCS B-Series Chassis with up to 80Gb of bandwidth per blade, not only can you reduce the physical foot print of your resources; you can do it with only a fraction of the cables and network ports needed for equivalent power and bandwidth to physical servers. A 42U server rack full of 2U physical servers would reduce down to just 18U with two slots to spare and can reduce your cabling and network port requirements by over 60%.




If your data center is suffering from a bulging waist line with no sign of an end to the overindulging of rack and network port space, contact your Keller Schroeder Sr. Account Manager to find out how the UCS diet can help you get lean and maintain.




Proactive Performance Management

– Gina Minnette, Sr. Account Manager


Do you or your IT staff frequently find yourself in a reactive role regarding your systems and network? Do you use technology as a competitive advantage in the marketplace and need assistance maintaining the proper health of your unique environment? Or could you simply just use an extra set of hands now and then?


Keller Schroeder can help!


As part of our suite of Touch Point Services, we are pleased to introduce a new set of offerings: PPM (Proactive Performance Management) Agreements. To help you avoid finding yourself in a reactive position with your infrastructure, let us assist you in taking a proactive approach by ensuring your systems are running as efficiently and effectively as possible. Keller Schroeder has PPM Agreements available for your Voice, Network, Virtualization, Storage, Systems and Security environments — all of which include Advanced TechLink features such as ticket prioritization, shared organizational view, and a Members Only forum.


While each agreement is customized to your specific business needs, a basic Systems PPM includes (but is not limited to):

  • Log review, capacity planning, and system assessment
  • Summary of systems status, including improvements recommendations and areas of concern
  • Installation of patches and upgrades for servers and workstations
  • Documentation of all systems covered in the agreement
  • Technical input on design and architecture decisions
  • Systems administration and maintenance for all Windows servers


    Keller Schroeder engineers assigned to your agreement will perform the tasks above at a scheduled, mutually agreeable time and with each PPM Agreement, the benefits are clear: less unplanned downtime and more productivity for you and your team.


    Contact your Keller Schroeder Account Manager for more information regarding Proactive Performance Management agreements.


  • “What we’ve got here…is a failure to communicate.” – VMI: Vendor Managed Inventory

    Vendor Managed Inventory Replenishment


    Have you been approached by one of your customers about starting a Vendor/Supplier Managed Inventory (VMI or SMI) relationship, or are you considering implementing one with your suppliers? While there can be benefits to both parties in a VMI relationship, there are pitfalls to avoid in order to establish a successful partnership.

    Potential benefits include:

    • Shorter lead time
    • Reduced inventories (for both parties)
    • Greater manufacturing efficiencies
    • Stronger relationship with your customer/supplier

    Potential issues include:

    • Accuracy of data: Without accurate and timely reporting of current inventory quantities and planned consumption (planning schedule), it is impossible for a supplier to keep their customer properly stocked with inventory. 
    • Exchanging data: Are both partners “on the same page” with the transmission methods (traditional EDI, XML, spreadsheets, proprietary formats, etc) and frequency? 
    • Managing the data: Depending on the number of products a supplier provides to a customer, the volume of data to be managed on a daily basis can be staggering – current inventory quantities, month’s worth of planned consumption data, min/max inventory levels, etc. As a supplier, are you going to add to your staff to manage this additional data, or will you be looking for a software solution to help your existing staff manage all of this new data? 
    • Consignment inventory: Some companies implementing VMI are also wanting to convert the inventory to consignment – meaning that they don’t own (or pay for) the goods until they pull them out of inventory and send them to the production floor (or sell them in the case of a final product). This relationship model adds more overhead to the supplier, in addition to the data exchange between the two parties.

    Properly implemented and managed, VMI can strengthen the relationship between a supplier and a customer, and provide tangible benefits to both parties. However, if not properly implemented, this relationship can cause issues for both parties, driving up costs, causing production downtime (due to a lack of available inventory), and can ultimately damage the relationship.

    Keller Schroeder can assist you with the process of planning and implementing a successful VMI relationship. Please contact Dan Ehrhart at (812) 474-6825 to discuss this topic in greater detail.


    “Run Forrest Run” – WAN Optimization


    Accelerating the Distributed Network


    As the work environments of organizations and employees become more distributed, the productivity of users in branch and remote offices becomes increasingly important to the success of the business as a whole. All users need fast, secure access to critical applications from wherever they are working; at headquarters, in branch offices or when traveling. Contradictory to this, for ease of management, enhanced security and compliance, the current trend is to consolidate applications and servers, physically or through virtualization, to fewer, centralized locations,. This simplifies server and application administration but can create performance problems for distributed users. In some cases, a centralized application results in a poorly performing application.

    Why?

    The problem is latency (the delay in responsiveness between systems); another measurement of a network that isn’t often talked about by telecoms and ISPs. That’s because it’s largely outside of their control. Beyond buying the fastest routers or laying more direct cable between locations, there is an inescapable fact: the speed of light. It takes 150ms – about 1/8th of a second – to go over a network from New York to Tokyo when the conditions are right. That doesn’t seem very long, but it’s over 100 times longer than the wait on a Local Area Network (LAN). That turns out to be important, as LAN application best-practice is to use thousands or tens of thousands round trips of small packets for every action. Even 1/8th of a second can add up to a significant delay when thousands of round trips are needed. What was fine at low latency (over the LAN), is often impossible at WAN latencies.

    Unaware of this problem, your employees still want to use LAN applications everywhere. They’re lightweight, flexible and easy for them to use – which results in higher productivity and lower support costs. On a LAN, they’re also fast. Ideally, there would be a way to enjoy LAN applications everywhere, regardless of the latency factor.

    Of course, we can’t ignore the impact of bandwidth completely. Applications such as email and file shares are now joined on the WAN by a new set of time-sensitive applications, such as Voice and Video. Accommodating those on the WAN, alongside the existing traffic, requires a way to address not only the inherent latency, but also manage bandwidth. The right user needs to access the right application without having to over-buy bandwidth for peak loads or frivolous traffic.

    Bluecoat‘s WAN Optimization Solutions accelerate performance to remote users to LAN-like speeds. In addition, policy based controls triage the use of network resources by matching users, applications and content with responsible use policies and your priorities.

    The following are business benefits of a WAN Optimization project:

    • Use the Internet instead of expensive private WAN links
    • Reduce costs by delaying the purchase of additional bandwidth
    • Enable key IT mandates, such as server consolidation, outsourcing and SOA
    • Remove redundant and inappropriate traffic to ensure availability at peak times
    • Increased user productivity by faster delivery of applications and data to remote users

    Bluecoat’s WAN Optimization Solutions are appliances to be deployed in the remote and headquarters offices. These appliances improve performance in several ways. By deploying WAN Optimization across the network, organizations can improve user performance while reducing costs. Server storage or consolidation can be centralized into the datacenter to reduce administrative expenses without negatively impacting the end user experience. WAN costs can be contained and possibly reduced as duplicate data is removed from the WAN connections. Branch users can become more productive no longer having to wait for applications and data.

    Contact your Keller Schroeder Senior Account Manager to learn more, see a demonstration, or discuss the benefits of implementing the Bluecoat WAN Optimization Solutions within your environment.


    The Blue Coats Are Coming!

    Blue Coat ProxySG Secure Gateway

    The Blue Coat ProxySG Secure Gateway (ProxySG) appliance can be configured to provide two different functions. In MACH5 mode, it operates as a WAN Optimization device. In Proxy mode, it functions as a Content-Caching Proxy Server, providing an additional layer of security for accessing the Internet.

    To “proxy” is defined as “authorized to act for another’. The ProxySG can be implemented in ‘forward-proxy’ mode where it is used as a client gateway to the Internet, or in ‘reverse-proxy’ mode where it provides the initial point of access to your web servers. In reverse-proxy mode, the ProxySG can provide services such as load-balancing and SSL offloading to enhance the security and availability of your web servers. For this discussion, however, we will focus on features provided by the forward-proxy mode of the ProxySG.

    Web browsers such as Internet Explorer and Firefox receive content from a web site and store that data temporarily in a folder on your computer. This temporary storage is called ‘cache’. If you go to that web site while the data is in cache, the browser can pull the data from there instead of going to the Internet to get it. This allows the web site to be displayed much faster. Unfortunately, this data is only available to the same person that initially accessed it since it is stored on their local computer. The content caching feature of the ProxySG changes that. Since it is a network device that is acting on behalf of your Internet users, the content is stored on the ProxySG. Therefore, the content from one user’s access is made available to all other authorized users. The next person that launches their browser and requests that same website will receive it from the cache on the ProxySG. The end result of this content-caching feature is that many websites can be stored on the ProxySG and made available to all of your users. This approach can help you use your Internet bandwidth much more efficiently.

    The ProxySG can support several web filtering solutions on-the-appliance, such as SmartFilter and Websense in addition to Blue Coat’s own BCWF (web filter). All web filtering solutions categorize web sites based upon a number of criteria such as key words, phrases, site language, links, and spyware. The BCWF database contains over 15 million websites published in more than 50 languages, and organized into 69 categories. Unlike some solutions that update the database daily (or less often), Blue Coat’s BCWF database updates these categories several times an hour. In addition, the BCWF web filter can be configured to provide dynamic protection with their Dynamic Real-Time Rating (DRTR) service that categorizes web sites ‘on-the-fly’. When running the BCWF web filter on the ProxySG, the powerful policy engine allows use of the categories in your defined access policies. For example, a policy could be written to block sites in the ‘Social Networking’ category during working hours, but allow access to those sites during the lunch hour.

    In today’s economy, online training is often preferred over traveling for a training course. If several people within your organization decide to view that online training at the same time, it could significantly impact your Internet connection due to high utilization. To help mitigate this situation, the ProxySG has a feature called Stream Splitting and Caching. First, if a video stream can be cached, it is cached on the ProxySG in a fashion similar to the web content mentioned previously. Secondly, the Stream Splitting feature is the ProxySG recognizing that the same video is being requested by another person and splits the stream internally to serve the content to both individuals directly from the ProxySG, instead of pulling the content from the Internet again.