critique or comment (with reputable source citation) on this
report ?
ASHEVILLE DISASTER RECOVERY
Executive Summary
A stable and prosperous city, Asheville in Western North
Carolina operates a Website which has information about the city
and can link to various online services. Asheville also created and
maintains a popular mobile app for citizen services. Jonathan
Feldman, the city’s chief information officer, who participated in
the Hurricane Katrina recovery, was concerned by what he found.
Asheville has a disaster recovery facility, but it’s only two
blocks away from the city’s main data center. Also, it can take
more than a day to switch from the main data center to the backup
servers in an emergency. Given that citizens rely on information
about city services during emergency situations, Feldman worried
that the city’s web servers would be offline for a long time,
especially those that support mobile applications. The cost of
operating a disaster recovery facility is expensive because it
needs to replicate the servers, storage devices, and software of
the regular data center. The additional facility also requires
reliable dual backup power system. Another factor is the facility
must be strengthened to withstand the forces of natural threats.
The recovery facility should be far enough away from the main data
center to be unaffected by the same disaster. However, it is a
challenge to convince city councils to spend large quantities of
money on computer equipment that might never be used.
Problem Statement
According to FEMA, “40 percent of businesses do not reopen
after a disaster and another 25 percent fail within a year.
Similarly, the Small Business Administration reports that over 90
percent of businesses fail within two years after being struck by a
disaster. Preparing for a disaster can make all the difference, and
it doesn’t need to be overly expensive or taxing operation”. Many
companies choose Asheville for the data center because Asheville is
one of America’s safest weather cities and it is far moved from
targets subject to man-made threats. Asheville operates a web site
with information about the city and link to various online
services. In addition, they create a popular mobile app for citizen
services requests. You can search for what you want to know through
this app. Travelers can use the software to learn about local
attractions, activities and hotels. However, the disaster recovery
facility is only two blocks away from the city’s main data center.
If a disaster occurs, the facility is likely to be affected and no
available server can use. This will lead to concerned or endangered
citizens that usually rely on the city’s informational
infrastructure.
Alternatives
When Jonathan Feldman did his review of Asheville’s
information infrastructure he established fixing the city’s
disaster recovery infrastructure mainly focused on their backup
data center. After reading up on the topic there are three
identified alternatives that Mr. Feldman could implement to remedy
the situation. These three alternatives include:
1. Fix Data Center Switchover time:
The current situation Asheville finds itself in is both data
centers are relatively
close which could be a problem depending on the disaster the
city may face. However, moving a data center is quite a costly
undertaking and sometimes getting upper administration to approve
the costly expenditures on infrastructure that may never be used
can be difficult. This alternative tries to get the city to focus
on fixing the switchover time of the backup data center in the
possibility that the moving of the data center is not feasible.
Although, keeping the data centers is not the ‘by the book’ way to
run a backup data center, sometimes a situation does not allow a
change to those circumstances. With the current setup the
switchover time in the event the main data center goes down is over
a day for the backup generator to pick up the services, which in
the event of a disaster is not ideal because a few hours can make a
difference, let alone over 24 hours. If the city invests in a
better system to run both data centers as mirrors of one another
the two can work together as load-balancing data centers. Through
this system it “directs end-users to the appropriate replica so
that the route to the replica is good from a network standpoint and
the replica is not overloaded” meaning that if one center fails the
surviving center can pick up the slack a lot more responsively than
if the current system is left in place as it takes over 24 hours
for the switchover (Leighton, 2016). If the two centers work
together if you lose one in a disaster then the data center that is
still running will start directing users to only itself as the
other center is unresponsive. With this alternative users of the
city trying to get information during the disaster would not even
know that a data center was lost as the backup would be much more
immediate and because of that the city would be in much better
shape for mass communication for updates on the situation.
2. Move Data Center into different geographical
location:
This alternative would duplicate the information stored at the
current data center
to a different geographical location. According to the
article, Everything You Need to Know About Data Centers or Mission
Critical Facilities, “A colocation facility is generally defined as
a shared data center space in which a business or an organization
rents space for servers and other computing hardware. Generally
speaking, the colocation provides the building, HVAC, power,
internet bandwidth, and physical security so customers can supply
and maintain their hardware” (Salarifard, 2019). There are two
approaches to this alternative that can be taken; buy facilities or
partner with a colocation service provider. Essentially, in this
aspect, the best choice would be for the data center to partner
with a colocation service partner. By doing so, the company saves
the costs of buying new facilities on their own in addition to a
plethora of other costs. The main benefit of colocation is to ease
the transition from one geographical location to the next. Having
this partnership certainly lessens the risk of losing or
compromising information as well. In the event of a disaster,
having additional facilities would come in handy. While one
location is recovering, the other location would be fully equipped
and prepared to handle the loss of the other facility. Another
benefit of splitting a data center is the increased ability to
serve a wider range of people. Since the data center will be in
more than one location, the users at those locations will largely
benefit. Ultimately, the city is able to increase its system
responsiveness and create a better experience for its users. Also,
colocation lessens the switchover time versus choosing to buy
facilities. The service partner would aid in the employment of the
facilities, the location and set up time frame. Edwards says the
colocation process can happen in just over two months. On the other
hand, there are a few risks associated with making such a big
decision; cost being the biggest. According to the article, Grow
Your Data Center with Colocation, the author shares that,
“Financial considerations may play the biggest role in colocation
decisions. Do you want to go to your board and ask for $50 million
in capex (capital expenditures) for another data center? The
alternative is to go to a provider and use opex (operating
expenses) and not have to spend money upfront.” (Edwards, 2012).
Edwards believes that colocation is an asset that needs to be
handled with care and expertise. The next big risk associated with
colocation is security, which is both virtual and physical. This
also includes backup infrastructure like power, cooling and fire
suppression. If handled improperly, the entire data system would be
at risk. Overall, choosing to move the data center into a different
geographical location is not the most feasible solution here
because of the risks associated with data colocation when compared
to other alternatives.
3- Utilizing Cloud Services:
This alternative would allow Asheville to lift some of its
existing services from
on-premise applications to applications running in the cloud
to take some of the worry away from personal infrastructure during
disasters. There are many options that Feldman could pursue within
Cloud Computing ranging from hosting their mobile app’s servers in
the cloud all the way to moving their entire backup data center
into the cloud. The benefits for using these cloud services all
depend on usage and what route Feldman would take. Looking at what
option would be best for Asheville leads to two models; a hybrid
model or a full cloud model. Using a hybrid model would consist of
Asheville’s data center staying where it is with the city moving
its mobile app and web servers to the cloud, for example into an
Amazon EC2 server instance. Researching articles on this topic
shows “the addition of persistent storage to EC2 is what makes
hosting a website on the service practical” which means that now
services ran on an EC2 server can be saved meaning now they are now
more reliable since they do not lose their information during
termination (Diginmotion, 2010). This would be a good option
because in the event of a disaster that affects the city the EC2
server would not be affected as it is running through the cloud
providers infrastructure and on top of that since the cloud is
accessible through the internet Mr. Feldman’s could still get
online through some means and push information to the city’s mobile
app and website to keep their citizens informed since the app and
website will still be up running on the cloud. The costs of the
city’s data centers would also be reduced because they could get
rid of the servers that currently run the app and website in their
own data centers. Another gain for Asheville would be they would no
longer have to worry about the website or app going down as any
cloud provider already has redundancy plans in place and if for
some reason one of the providers data centers goes down the slack
is immediately picked up by a different center. A con with this
hybrid model is the sensitive information hosted in the city’s data
centers is still vulnerable to natural disasters since both data
centers could go down since they are so close in proximity. The
other model Mr. Feldman could implement would be a full cloud model
consisting of moving all three services (web servers, mobile app
servers, and the backup data center) into the cloud. One pro of
this model includes the city would not have to plan for hardware
refreshes every 3 years as that is the average lifespan of
commercial servers which can be costly. Another benefit is storage
on the cloud would be much less costly than providing their own
storage on site as Amazon’s S3 mass storage service runs at $0.023
per gigabytes which even running up to 50 terabytes on S3 would
only cost around $1,100 each month which is cheaper than on site
storage would cost and you also gain access to Amazon’s high
bandwidth as an additional plus. One of the main benefits that Mr.
Feldman would consider when looking at a full cloud model is that
the city’s entire data information system would be backed up in the
cloud and through any disaster the city could see all of the data
would be preserved and the website and mobile app would continue to
run passing information to the city residents. Also, as long as the
proper steps are taken during the planning phase of moving the
services to the cloud Mr. Feldman could reduce the overall cost of
maintaining the backups of the services since cloud computing cost
is based on usage and the mobile app and web servers would not have
to spin up their resources unless a disaster hit the city and they
were needed. The main cons for a full cloud model would be if any
privacy laws prohibit certain information being stored on cloud
services, if any data privacy laws change the city may have to move
some of the data back out of the cloud, vendor lock-in with certain
cloud providers, or the possibility if proper planning is not taken
during the migration process the cost of these backup services
could get out of hand.
Conclusion
Ultimately, we agreed to suggest Mr. Feldman to convince
Asheville’s city council to spend money on Cloud Services.
Alternative 3 would serve as the best option for the company. One
main reason is this can reduce the costs associated with the backup
center’s infrastructure. Cloud computing can range from servers
hosting mobile applications to moving entire backup data centers to
the cloud. The city’s entire data information would be backed up in
the cloud and it would not be destroyed or lost through any
disaster. Websites and mobile app would continue to run passing
information to the citizens. Feldman doesn’t have to worry that the
city’s web servers would be offline for a long time Also, Feldman
can select and adopt hybrid model or full cloud model according to
the situation of the city and what the board approves pending
compliance regulations of the city and state.