Recent Tweets
join our mailing list
* indicates required

About Ambleside

Ambleside Logic is led by Aaron Rosenbaum. Father of 3, Programming since 7, DevOps since 11 (hacking RSTS), exIngres, exCTP, exCohera. Sold two companies to Oracle, one to HP. Research + Strategy for NoSQL/BigData ecosystem implementors, vendors and investors.

Tuesday
Jul052011

Faster BI time-to-value in 2012

A few ideas for your 2012 project portfolio:

Sponsor an Analytics as a Service project

Find a project with rapid business value and a small number of end-users.  This would be a simple project to do internally.  Don’t.  Find an outside provider to supply the whole stack - analytics, implementation process, software stack - Merced in Sales Performance, RetailDriver or AcceleratedAnalytics for POS, etc.  Being able to manage entitlements, dats security, and integration while utilizing an outside analytics application provider is a key driver for increasing agility.


Use Hadoop (mapReduce) and Python for an Extract/Load project

Yes, Hadoop can enable incredibly large scale big-data projects.  But new technology and new business solution at the same time is not a good equation.  Try piloting Hadoop on a plain vanilla

Click to read more ...

Friday
Jun242011

Blow-up the data center! Which industries should go 100% cloud.

What top 5 industries are ripe for blowing up the data center?

They should have:

  • Distributed user base (already dependent on network for operations)
  • Lack of differentiation on IT
  • Business processes tending towards standard because of regulation or custom
  • Advantage to move to highly distributed client devices

#1 - Primary Education, Small and Community Colleges.  

Employees and "customers" are highly distributed.  There are lots of SaaS providers already in this space.  The user base is much less elastic than other SaaS markets but the advantages in system administration and client admin should overcome the lack of savings there.  I do think this is ripe at the moment because of pricing pressures.  Processes are more complicated than simple service businesses or Restaurants but still, not that bad.

Click to read more ...

Thursday
Jun232011

Product Marketing Vision in Virtualization Stack Providers

Give me a solution to migrate existing app servers to a new virtualized infrastructure...Tell me your point of view and solution.  With billions of dollars being spent on solutions marketing, cloud management, you would think this would be prevalent with the issue being which person to listen too...Nope. No substance, lots of fluff.

I went looking for something fairly straightforward - which IT equipment vendors had documented solutions for a virtualized stack - of practically any sort - without buying a vBlock, flexPod or vmTank, etc. There are lots of  choices. I want to see how the vendors themselves see the architecture fitting together - at all levels.

I want to see:

 

  • The big vision - why do I care, who is this for, what benefits do I get by following their advice
  • A detailed design guide that walks me through key components, sizing, capabilities, technical approach
  • A deployment document that gives me exact model numbers for the validated design, configuration scripts (even if they repeat the manuals), and decision trees for configuration decisions.
  • Clear nomenclature and dependencies so when I call support and say "I did it like doc X01234 said to do it" they know what I'm talking about (or at least someone will.)
  • Let me find the content, somehow, on your website...Building virtual/converged/unified infrastructure is only the hottest thing to hit the IT field since App/Web servers - you'd think it would be upfront...

Click to read more ...

Wednesday
Jun222011

Converged Infrastructure for the $10M IT department

$10 million dollars is a lot of money. It's a decent budget for a non-financial industry IT department supporting $300M in revenue in retail, services or even some manufacturing industries.  With that budget though, you probably have to support quite a few mission critical legacy applications, certainly SAP or Oracle ERP, lots of MSFT licenses, several thousand desktops.  Taking care of the data center running the mission critical applications is a team of 5-10 really smart dedicated admins supporting everything the last 3 CIO's bought in the previous decade.

In my opinion, converged infastructure with hybrid cloud is the only way out of the current mess most departments are in.  Why not a move to public cloud and consumer apps?  Because typically, there aren't resources to deploy them. 

Example:  You have a mostly working Business Objects reporting infrastructure with 50 really active users.  Of those 50 users, 5 know how to revise and edit reports. The resources necessary to transition all the knowledge in those reports over to another platform just isn't there.  But, the data rates keep rising and it's hard to keep up.  This is a typical challenge. 

My definition of a converged architecture:

  • Physical dependencies are removed from the infrastructure
  • Hetrogenous, proprietary and legacy protocols are encapsulated in fairly homogenous protocols
  • Previously application specific maintenance operations can be applied on a more horizontal basis (especially disaster recovery.)
  • More horsepower can be deployed to different applications without over provisioning

Click to read more ...

Wednesday
Jun222011

Small Business Build-Out Storage

I've been looking at the smaller build-out storage options for a project I'm doing.  I've never seen such public bitchiness between the tech marketing teams as in this market.

The future for this company is a converged architecture - app infrastructure as an appliance.  VNE and vBlock are fine ideas but hetrogeneous support up and down, not specific storage or app servers, is likely to win the most market.

With that in mind, here are the requirements.

This project requires something that can get going quickly without running out of steam as they grow.  Like many infastructures these days, there is no need for anything that scales massively for a single query set.  But they don't want to roll their own solution entirely at the app level either. Storage sizing is estimated around 4-6Gb for main tables, 20GB for other data.  That means something with a 50GB-100GB capacity.

 

  • Market leader is important - resale value, ability to purchase used, market availability of expertise, etc.
  • Cloud-storage tiering for off-site and scale-out.
  • Really easy to use management tools - no dedicated storage staffing, no classes.
  • Really tight virtualization integration - back-up/DR/management - no science projects.
  • Total cost is important but CapX better than people as long as disc upgrades don't cost arm + leg (do any of these support BYOD.)
  • Open standards with published HCLs for hosts + host adapters.

Click to read more ...