O
perating
S
ystems
Companies that rely on specific Linux com-
ponents in the relevant branches either pro-
vide their own personnel to work in the open
source communities or become members in
the relevant communities of interest such as
the Linux Foundation in the USA. In this man-
ner, the beneficiaries of the software participate
in its development, and a classic balance of give
and take is generated in the community, a pro-
cess manifested in the implementation of real-
time Linux kernel features.
In the beginning, companies such as IBM and
Red Hat shouldered a major portion of the
costs in a concerted effort because real-time
Linux was sorely needed for customer projects.
Their requirements have now largely been ful-
filled. Nonetheless, the non-existing mainline
consolidation of the remaining components
does not constitute a disadvantage since the
systems can be individually addressed by a
small specialist team using a separate busi-
ness model. The corresponding product by
Red Hat is called MRG (Messaging/Realtime/
Grid); it contains a real-time Linux kernel
yet is only available for x86 systems (32- and
64-bit) and is only certified for a relatively
small number of hardware platforms.
The use of Linux within industry is a differ-
ent matter altogether: real-time features and
other branch specific extensions are required
in a variety of architectures and in hundreds
of different systems. Reliable and sustainable
development is only feasible if the missing
components are incorporated into mainline
Linux. Linus Torvalds has affirmed this repeat-
edly, with the proviso that additional com-
ponents must be incorporated by a team of
appropriately qualified kernel developers. He
especially pointed out that these developers
must also be available later on when real-time
components need to be adapted to new kernel
features and maintained.
The Open Source Automation Development
Lab (OSADL), founded at the end of 2005
for this purpose, has experienced consistent
growth and functions as a nexus for a signif-
icant number of industrial companies. The
available funds of OSADL have been used for
a while in quality assurance of the real-time
Linux kernel. When in summer 2014 is was
decided to cease further work on the real-time
patches because of lack of funding, OSADL
started a campaign among it members and
other interested parties to raise the required
funds. Fortunately, these activities were suc-
cessful, and a contract was signed to ensure
development, maintenance and bug fixing of
the real-time patches. The work resumed in
January 2015 with Linux kernel patch level
3.18 and will – as before – continue with
even-numbered patch levels.
This certainly is an important achievement
that is very much applauded by industrial and
other users of real-time Linux. It must, how-
ever, not obscure the fact that the second step,
i.e. the complete mainline consolidation of the
real-time patches remains the final goal. This
is particularly important, since mainline con-
solidation is the only way, at least in the long
run, to guarantee the continuous availability
and industrial quality of real-time Linux (fig-
ure 2). It is easily conceivable that OSADL will
be able to shoulder this remaining work with-
out requiring external funding. Until then,
however, OSADL is urgently appealing to all
users of real-time Linux to join the common
effort to make this final feat happen.
n
Figure 2. Time course of the efforts to provide and maintain Linux real-time and the positive
effect of complete up-streaming
