Monday, August 31, 2009

Sample argument for a Blade Center

Here is a sample discussion paper I wrote to assist a First Nation Community in evaluating a blade center approach to infrastructure.  The name of the community and identifying details have been altered or removed.

What the following post highlights is that the model changes with blades and virtualization to the positive.  However, an organization’s thinking must change to achieve these benefits.

Issue

The Community administration’s is looking at strategies to upgrade their server infrastructure.

Background

There are currently 11 servers within the community at different stages of their lifecycle. Many are end of life and present a critical risk in the event of failure. The traditional method of creating a business case for each server is both time consuming and may be preventing future opportunities. This briefing note will attempt to demonstrate another alternative.

How does it work today?

In today’s environment when a server becomes end of life the organization is faced with several challenges. The first challenge is to understand the role of the server. Is the role of the server mission critical? Some of the servers today are considered mission critical. The more mission critical a server is the shorter its outages can be. These servers perform functions like email, accounting, grades, Firewall, printing, file sharing etc. Each and every one has its role to play and it may or may not be able to be offline. Only the organization can make that assessment.

Once this assessment is done, then the organization has to determine how big and expandable does the new server have to be to last until that platform is end of life. The recommended recycle time for critical servers is three years. If you follow this approach the first years of the server implementation the server is running well below optimum, the second year it starts to realize full performance and somewhere in its third year the server will start to become strained from overuse. This is a typical scenario and is repeated for each and every server. These ebbs and flows are constant in the industry.

To address this static hardware replacement strategy the industry is now deploying what is called blade centers. These blade centers take the best of today’s hardware and packages them so they can fulfill the needs of many servers on one hardware configuration. This virtualization of physical servers to many logical servers is well demonstrated and proven in the industry. The technology has been proven and it is being used right here on Manitoulin Island at Manitoulin Transport. They were able to consolidate over 100 pieces of equipment down to 10 pieces of equipment while at the same time expanding their services offerings.

Costing Sample

The following table demonstrate the costing model difference between upgrading 11 servers independently verses using one blade center. The blade center described is capable of running 18 servers of the type and model being utilized by the community currently. I have chosen 18 due to the fact more servers are actually being recommended by all the major vendors today. It will also allow the community to build testing environments without having to purchase new hardware.

blade costs

As you can see the second year of the model the cost structure dramatically switches in the favour of blade centers. If you factor in additional servers then the case for blade center is proven sooner. The third and out years are even more favourable to blade centers.

Benefits of Blade Centers

  1. Centralized control and management
  2. Cost savings
  3. Centralized Backup
  4. Back and recovery per server is consistent and resilient
  5. Annual incremental increases verses sporadic larger capital costs.
  6. Autonomous building of servers
  7. Integrated security
  8. Improved network performance and resilience
  9. Reduced support costs
  10. Small form factor
  11. Reduced energy costs
  12. Reduced heating and air conditioning requirements
  13. Industry adopted
  14. Many vendors to support the platform and its implementations
  15. Increase the number of servers with minimal or no additional costs
  16. Increased flexibility for your technology rollouts
  17. Hardware independence from your operating systems. Bigger benefit in the Microsoft world.

Burdens of Blade Centers

  1. New skills are required to maintain the platform
  2. Software to build the platform will need to be purchased
  3. Migration project will need to be undertaken to port the existing 11 servers to a common platform
  4. Organizational policy change to support central housing and shared resources
  5. Governance structure for support and backup rotations
  6. Annual operational dollars will need to allocated to sustain the platform

Conclusion

The cost of hardware is plummeting but the challenge of effective use of hardware is ongoing. By adopting a blade center strategy, the community can reduce costs, increase effectiveness and provide a long term platform for the entire community. Each of the current 11 servers is important and their ongoing support and functions are critical to the overall community. A proactive and long term view to their deployment will have a positive impact for the community.

Next Steps

  1. Agree to the strategy
  2. Build a requirements list for the platform
  3. Build a RFP
  4. Source the platform
  5. Build the migration plan
  6. Perform the migration

Monday, August 17, 2009

The challenge of Voice over IP

There are lots of discussions regarding Voice over IP (VoIP).  They typically are centered on the technical merits or shortcomings.  This post will attempt to address some of these issues but more importantly address the business discussions that should be transpiring.

First, my biggest pet peeve is for the vast majority of the naysayers who claim with great flair and passion that “VoIP is not ready for the real world”.  Ok, check the sources of these claims.  Most likely is the the big telecomm groups. Why?  Simple, they are trying to keep as much money in their pockets as long as possible.

Every technology has growing pains, the telco’s have had a hundred years to better the experience. Recent failures of the big telcos show they didn’t see the real business problem but are still bragging on the technical wonders of what is in their labs. 

This graphic shows the timeline from the old copper days, to digital phone (the said saviour of the business) to Internet Protocol (IP) phones.

VoIP

The point here is that there has been a path to something different over time and with increased functions.

The large phone carriers have been doing VoIP at their interconnection points since the late 90s.  These are the same carriers that say VoIP is not ready.   My challenge to them has always been, then what are you doing to get it ready?  Their users have spoken and answered in droves by saying “not much”.

I personally have been doing VoIP at the trunk level for my enterprise phone switches since 1999 over Framerelay, yes Framerelay.  So much for the engineering specifications of today’s massive network requirements.  Which leads me to the real challenge of the convergence.

If VoIP is going to work it has to travel on a good data network.  Seems simple but here is where the real cultures clash.  The clash is not in technology but in approaches to solving issues.  The 100 plus years experience of telephone thinking is move slow (read glacier speed) and over build.  The premise always was, we have lots of money and time.  Along with this the customers really had no choice.  If you wanted voice mail and auto attendants the choices were limiting, confusing and very expensive.  This voice culture rewards safety and engineering prowess.  In the traditional phone world the engineer dictates the experience, hence the horrible user interaction points.  The only saving grace was it worked predictably and reliably.

The culture of the data teams is get it in and fix it later.  This is a plaque that is ruining the industry.  Poor network design along with sloppy implementation coupled with cavalier change management policies is their legacy.  This culture leads to some of the vast security breaches that occur today.  The network needs to be designed and implemented properly to really gain the value of VoIP.  Some of benefits of the data network engineers is one of adaptation and creativity.  This is a good trait to have as long as it is well managed and lead.  In this new world the network is king so let’s not make the village idiot the king.

On the same note, let’s not make the telco trolls the masters of all.  This is way too dangerous and counter productive.

The Internet and the internet protocols are here to stay.  Many good people have built worthy networks capable of running VoIP simply and effectively.  These people have used the true value of voice engineering design principals with the creative approaches and techniques of the data network arena.

The business case for VoIP is simple, reduce the moving parts and players and you gain the competitive advantage.  It is easier to find network experts than phone experts.  The technology being created for the VoIP space is being done by creative people solving real business problem.  They are also using modern business models that are more cost effective and give the businesses real options to solving a vast range of business issues.

We have currently converted 5 of our major phone sites to VoIP in support of 600 users.  Phone updates are done by the PC techs using the same tools they use to setup a new Windows users.  The days of the black arts being managed by the Merlins are coming to an end and that is a good thing.

The kingdom should be managed by the king on behalf of the kingdom for the benefit of all.  VoIP is an effective business tool to get there.  Yes there will be scary witches, there always are when you are implementing big change.  The alternative is to continue to pay too much, get too little and be locked into old rigid thinking.   Modern companies have fought through the technical issues and now are reaping the benefits of mobility and true business control of their phone infrastructure to gain business advantage.

Some benefits to the business are;

  • Ease of Moves, add, changes (MAC)
    • Can be done by PC Techs
  • Mobility of workers without support calls
    • Worker can change offices without have to call for new cabling or setup
  • Everything is a data drop CAT5e and CAT6
    • No more dual copper connections
      • voice CAT3
      • data CAT5e and CAT6
  • lower licensing
  • Cost based on users not systems and users
  • More competitors for different options within same environment
    • phone systems
    • voice mail
    • call recording
    • reporting
    • phone sets
  • incremental user costs for new users
    • Can purchase as needed instead of in bulk
  • Session Initiated Protocol (SIP) trunking for dial tone
    • Very inexpensive and very reliable regardless of what you hear from your vendors.  Talk to actual users of it
    • Significant business continuity benefits and toll reduction over traditional PRI setups

Some draw backs and cautions to VoIP

  • The data network must be checked for compliance
    • DO NOT underestimate this.  Just because it is working does not make it workable for VoIP
  • Tough to find a vendor that is good at both voice and data
  • Tough to get the voice and data teams to behave nicely together
    • expect causalities if you have mature teams.  This is not always a bad thing
    • They need to learn from each other without coming to an impasse
  • Watch out for the hidden licenses.  Read the fine print and ask lots of questions.
  • Every thing is server based and with all servers then need to be maintained in a proactive way.
  • Make sure your core data network is near any voice recording software.  Voice recording in the new VoIP world is very easy and let unmonitored will kill your data network with traffic.
  • Make sure you are solving business problems when you purchase some of the optional offerings.  This is where their money grabs come from.  Do not be afraid of mixing your vendors.  Is this new world is it very easy to do because of established standards.

Test from Windows Live Writer

This is test to see how Windows Live Writer works.

kendo

  • This is me when I prep for my Sensei
  • Be afraid but do it anyway
  • Train today to better then you were yesterday to defeat the enemy of tomorrow

I have to admit the tool is really kewl. Setup is a little odd and it does not read the Internet settings from the O/S or IE. For a Microsoft product that is odd.

I will be using this tool on a regular basis. Thanks Frank for showing it to me.

Thursday, August 13, 2009

Thoughts on operating systems

Here is a link to a friend of mine. His thoughts on operating systems is well worth the read.

It is brilliant in its clarity even if its implementation is hard work. Doing things right is rarely easy.