[vcf-midatlantic] Great vintage computer article.

[Dave McGuire]

   Most companies lost money in the face of downtime.  The specifics 
vary from company to company, and analyzing/weighing the risks is a part 
of the process.  This is what Bill Degnan and I just exchanged email 
about..."people who know what they're doing."  Sadly a lot of these 
cloud migrations are happening without the benefit of people who know 
what they're doing.

   And another problem is that those downtime averages aren't etched in 
stone, they vary over time.  Then you're faced with weighing the expense 
and risk of a different provider, in addition to the expense and risk of 
the actual move to that new provider.  The company my main contract is 
with has been in the process of moving their services from Rackspace to 
AWS for about three months now.  It has been more expensive than anyone 
ever thought it would be.  I'm glad I don't work on that side of the 
company!

   But as Bill said, and you have supported, it's all about knowing what 
you're doing and figuring out what's the best approach for your company. 
  If YOU (or Bill Degnan) as a technical professional with subject 
matter expertise decide that it's the best path, go for it.  But most of 
these companies just jump blindly, issue a directive to "MOVE TO THE 
CLOUD" and turn a blind eye to the risks because they see dollar signs.

                -Dave

On 04/19/2017 01:09 PM, Drew Notarnicola via vcf-midatlantic wrote:
> That's why you need to be careful what cloud provider you choose, and look
> at their average downtime.  If you're not trading stocks and a few minutes
> of downtime won't lose you much money, then you can go with a provider that
> only sees a few minutes of downtime a year.
> You also need to look at what your reliability is going to be by keeping
> your servers in house, as well as how much it will cost.  It's very
> possible that a cloud service can provide you better reliability,
> redundancy, and immunity to natural disasters than you can get in house, at
> least without spending a good deal of money.  That's when you go with the
> cloud.
>
> On Wed, Apr 19, 2017 at 12:24 PM, Dan Roganti via vcf-midatlantic <
> vcf-midatlantic at lists.vintagecomputerfederation.org> wrote:
>
>> On Wed, Apr 19, 2017 at 11:58 AM, Dean Notarnicola via vcf-midatlantic <
>> vcf-midatlantic at lists.vintagecomputerfederation.org> wrote:
>>
>>> I have to agree with Bill. Given the proper use case, cloud (true cloud,
>>> which is not just "someone else's computer", a common misconception) can
>> be
>>> an advantage. Moving some of our compute and storage resources to the
>> cloud
>>> allows us more agility to react to changes and liberates our limited IT
>>> resources to engage in higher value strategic work.
>>
>>
>>
>> ​I don't have a argument here, other than to say their reliability factor
>> is a total crap
>> ​When you have 1,000's of businesses that rely on commerce and this all
>> travels thru these cloud computing providers,
>> And the majority of commerce is done online these days
>> Yet they are a single point of failure - when Amazon goes down because some
>> nitwit type in the wring command line - just recently
>> POOF went all business for every one of those companies -- in the Millions
>> !!
>> When I worked in engineering at Bell Labs for ATT and Lucent,
>> You are >>>>required<<< to implement Fault tolerance mechanisms into your
>> design down to the bare metal.
>> This was a long die-hard tradition from the days when the majority of
>> commerce was done over the telephone - ie.before the Internet.
>> Availability was the competitive factor in the 5ESS and previous models,
>> also the BWM Network Transport systems for the backbone,
>> so it's not just a FCC requirement.
>> Networks have no regulations for Availabilty
>> Dan
>>


-- 
Dave McGuire, AK4HZ
New Kensington, PA