“We are outsource Big Data to Oracle”: Q&A with Devops Borat
JAXenter speaks exclusively with visionary twitter pundit @DEVOPS_BORAT on cloud downtime, outsourcing big data and (of course) devops.
the thousands of pundits voicing their thoughts on Twitter,
stood out with his strong views and piercing insight on startup
culture and emerging technologies.
Borat has managed to attract over 30,000 followers on Twitter so
far – more than Patrick
Debois, who coined the term ‘devops’ – and serve as inspiration
for several other developer Borats including @cloud_borat, @BigDataBorat and @AgileBorat.
Life of devops in 6 word: “I thought it was staging
— DevOps Borat (@DEVOPS_BORAT) March
OpenStack is Esperanto of cloud. Same adoption.
— DevOps Borat (@DEVOPS_BORAT) July
In devops is not correct ‘shit hit fan’ but ‘shit hit load
— DevOps Borat (@DEVOPS_BORAT) August
Is not site outage if nobody is tweet of it.
— DevOps Borat (@DEVOPS_BORAT) November
We wanted to find out more about the man behind these
140-character nuggets of wisdom. And before you ask: yes, this is
the real deal.
JAX: How did you get into Devops?
Did you start with dev or ops?
Borat: You not get into devops, devops is get
into you. Is not matter if you start by dev or ops because you are
never know what devops is if you are never awake by pager at 3 in
morning. Dev with not ops is as knowledge with not wisdom or as
theory with not practice like we are teach in Soviet Union. Ops
with not dev is also call junior sysadmin.
JAX: Can you describe an
average day for you? What languages and tools do you work
Borat: Is not such thing as average day in devops.
Every day is bring own fire which need of extinguish. Is just
matter of small fire or big fire. Small fire is usual on dev side
and is call bug. Big fire is usual on ops side and is call site
outage. Is why is rare of find ops older of 45 years because they
can never able sleep sufficient.
“Language” is strong word in devops. We are usual work with
shell script in for loop. Some time we are pretend of use script
language as Ruby but we not know what we do 90% of time. We prefer
tool write by other devops which we usual try of install for 3 day
then go on IRC and find solution in 5 minute.
JAX: A big part of Devops is
continuous delivery. How often do you deploy a new
Borat: In startup we are force new engineer for
deploy own code in first day at work. Is why we have 5 ones uptime
for web site. Also continuous delivery is concept of 3-4 year ago
is get old by now. In startup we have button on home page which is
allow every register user for deploy latest build + patch of them
own if they like. We not allow anonymous user for security purpose.
We are find this method is enhance user experience and user trust
in our product (we are only have 1 user so far).
JAX: How is your company adapting
to the rise of smartphones and tablets?
Borat: We make bold decision of have product only
available on smartphone and tablet which in real world is mean just
iPhone and iPad. As result we double user base from 1 to 2 which is
great success for mobile compute. Revenue is of negligible because
nobody is click on ad but we are go after market share and make big
just as Color.
JAX: How do you manage your Big
Data? And do you have a favourite NoSQL database?
Borat: In startup we are outsource Big Data to
Oracle. In return we are give them all profit from web site but we
get small cut of it back plus we can able find way of make more
money because of Big Data Analytic they are provide. Is good
strategy, I recommend.
All NoSQL database are look very good on single server and on
vendor press release but nothing is work after 6 month of real
world usage. Is when you are need of hire creator of NoSQL database
which fortunate is not hard because all creator already have
consulting company by now. Is cost lot of money so I recommend
follow same strategy as Big Data outsource.
JAX: AWS and Google App Engine
have both seen
high-profile outages recently, bringing sites down with them.
How do you protect your systems from downtime?
Borat: In startup we are not believe in downtime.
We are call it planned maintenance. As many expert on Twitter are
say, cloud is offer infinite scalable and zero cost but can not
able offer 100% uptime yet so we prefer use own datacenter. In
actual we are use N+1 datacenter where N is current zero but we
have big plan of change N to positive number in near future.
JAX: Finally, how do you see the
industry changing over the next year?
Borat: My top 3 prediction for 2013 are: 1) Number
of expert on Twitter is get double. 2) Cloud is able of guarantee
100% uptime with not extra of work from devops. 3) 2 out of 3
prediction from expert on Twitter are prove of wrong.
You can follow Borat on Twitter at @DEVOPS_BORAT. No, we still
don’t know who’s behind the account.
Image by Greg