In The Cloud?

We’ve been hearing about the promise of cloud computing for some
time. There are finally enough companies that have used the cloud to
have experienced the reality of cloud computing and learned some
interesting lessons.

So, recently, at its March monthly dinner meeting, the local chapter of the national association of CIOs (SIM) had a panel of IT executives discuss the migration to the cloud:

  • Len Peters, the University Chief Information Officer and Associate Vice President at Yale University
  • Larry Biagini recently retired as Vice President and Chief Technology Officer of GE
  • Jeff Pinals, Senior IT Manager for Enterprise Financial & HR Applications at XL Catlin
  • John Hill, COO of Virtustream, with responsibility for Cloud Platform Delivery and Global Data Center Operations.

I
moderated the panel.  Since there’s been so much written on this
subject, I’ll just focus on three revealing, yet not widely reported,
insights.

Security

There has been obvious
concern about security in the cloud, especially when a large amount of
data is held off premises.  Larry Biagini pointed out that those
security issues are already shared with enterprises that do not use the
cloud.

The old security moat around the enterprise is not a
modern defense in a world in which all computers are effectively
connected to each other and both employees and even trusted customers
are executing transactions through their personal devices.  A more
intelligent approach to security prepares the enterprise for its
migration to the cloud – whether it be the public cloud or a cloud that
someone thinks is private or hybrid.

image

The Cloud Project

Some panelists described the
process they used to select a cloud vendor and migrate to cloud
computing.  The tendency was initially to think that the whole story was
about following the usual steps in any IT project.  But Jeff Pinals
pointed out that migrating to the cloud is more than just another IT
project.  A good example is understanding how cloud computing will may
have an uncomfortable challenge from the organization’s culture – and
planning to address that issue.  Specifically, even in or perhaps
especially in companies with the best IT shops, non-IT managers are used
to a high degree of flexibility and accommodation to all sorts of
customizations.  That’s less likely to happen with cloud computing where
a SaaS (software-as-a-service) vendor cannot efficiently run the
operation by being so accommodating.

Return From The Cloud

Although
cloud computing is still a new experience for some companies, already
the question has been raised as to where this leaves an enterprise once
they’ve made the move.  The issue was highlighted by the news just
before this panel spoke that, after using Amazon Web services since it
started many years ago, Dropbox was leaving the Amazon cloud and
creating its own network and data centers.  See, for example, “Why Dropbox dropped Amazon’s cloud
published the day of our meeting.  It’s worth noting that even with the
large resources and technical talent of Dropbox, it took them more than
two years to make this re-migration from the cloud.

The panelists
indicated that there may be several reasons why moving or dropping out
of some other company’s cloud service would be desirable.  Perhaps it is
a competitor or potential competitor.  Perhaps its service wasn’t what
was expected and the decision makers were so burned by the experience
that cloud computing is off the table for now.

In Dropbox’s case,
perhaps the company is just sizable enough that the value added and
extra cost of using a cloud computing vendor no longer made financial
sense.

Whatever the reasons, after a couple of years, an
enterprise’s IT staff will also have migrated to a different set of
skills when someone else is handling the data center and related
operations.  The panelists noted that loss of data center skills may be
irreversible, at worst, or cost an enormous amount of money to rebuild,
at best.

John Hill ended this discussion that the move to cloud
computing requires a change in orientation about this loss.  Referring
to another utility we take for granted, he asked: Do you generate your
own electricity? Do you know how?

We need to realize that the
benefits of cloud computing have consequences.  Trying to return from
the migration is a bit like coming back out of the real clouds without a
parachute 🙂

image

© 2016 Norman Jacknis, All Rights Reserved

[http://njacknis.tumblr.com/post/141898219466/in-the-cloud]

Why Is Government So Slow To Collaborate On Software?

Open source software is free and is often developed by collaborations
of volunteer programmers around the world, along with staff at big
companies who find it in their common interest that this software be
maintained.

Open source software has been an enormous success.
Most of the web services on the Internet are delivered by Apache
software on the Linux platform – both open source projects. Indeed,
there are over 25 million open source projects listed on the open source
directory of GitHub.

Despite the billions of dollars spent by the
public sector for software each year, the public sector’s share of
those open source projects has been much smaller than its share of the
overall economy. This small commitment to open source is not what you’d
expect considering the situation of the public sector compared to the
private sector.

First, unlike private companies, governments
cannot really consider their software to be proprietary and a key
strategic advantage in competing with others. Taxpayers have already
paid for the software and, like much else in government, it is supposed
to be open and available (unless security or privacy is at stake).

Second,
there are many programmers who would help to build and maintain public
sector software out of civic spirit and a recognition that they too
benefit as citizens.

Code For America
has been an outstanding example, but not the only example, of the
willingness of software developers to help even local governments. Ben Balter has been a champion for public sector open source as well. The GovCode
website proclaims:

“Code for your country! We believe in a government
of the hackers, for the hackers and by the hackers”.

[That’s hackers in
the good sense of creative, expert programmers.]

Third, open
source software is often much less expensive than commercial,
proprietary products and just as often better quality. This is
especially true in the public sector where many of the companies
providing products for various special purposes – like jail management or health –
aren’t very big themselves and thus their products are often constrained
or sometimes buggy.

It’s hard to come by the total
number of software developers in governments or total expenditures for
software development – or even packaged software where open source might
substitute. But considering how many government agencies there are, it
is likely that government at all levels has a lot more developers in total than these
companies.

Of course, in fits and starts, open source has come to government, with 2012-2013 a recent peak of interest.

The US Defense Department issued guidelines
for use of open source in that agency. And the US Federal government
has, officially, been an advocate for open source software – at least
within itself. It is not easy to determine what percentage of software
the Feds have actually shifted from proprietary to open source.

image

Also, a few years ago, the Open Source Institute
dedicated itself to

“promote the development and implementation of open
source software solutions within U.S. Federal, state, and local
government agencies.”

Alas, currently it seems to be in a quiescent
state.

More recently, Github has created a platform for public sector sharing. But much of what is being shared is data or best practices, not software.

It’s
worth noting that, despite the efforts of the former DOD CIO and groups
within the General Services Administration, like 18F, too little of that movement
has been the result of leadership from within the government. Most has
come from outsiders offering to help.

Most significant is this
nuance: even when there is interest, government policies on open source
are focused on using open source software developed elsewhere, like
Linux, and not necessarily contributing to that software or creating new
open source software.

A key missing element in the failure to
make open source development the standard approach in the public sector is lack of
collaboration among governments. This is hard to understand in an era
when public agencies are strapped for cash. Each public agency may have
only a small software budget. However, pooling their financial and human
resources will achieve a scale that could allow for the creation of
good, feature-rich software for all of them.

This collaboration
need not stop at national borders. For example, the requirements for
software to manage the vaccination of a population are very much
consistent in the public health agencies of many countries. Moreover,
several trends should make collaboration easier – the shift to cloud
computing, new forms of communications that can ease discussions among
developers and the decreasing cost of building software compared to
years past.

It is now possible for government technologists, with
support from citizens, to truly scale up their open source software
development efforts.

So what are public officials, especially
public sector CIOs, waiting for?

If you have the answer or, better yet,
suggestions on how to make public sector open source more widely
adopted, please let us all know.

© 2015 Norman Jacknis, All Rights Reserved

[http://njacknis.tumblr.com/post/133033805117/why-is-government-so-slow-to-collaborate-on]