FreeBSD Foundation Newsletter, July 31, 2012
In this Edition:
- Letter From the President
- Fundraising Update
- Development Project Updates
- NAND Flash Support
- Growing Filesystems Online
- IPv6 Performance Analysis
- Capsicum Component Framework
- iSCSI Scoping
- Conference Updates
- Research Highlights
- 2012 Grant and Travel Grant Recipients
- A-Team Systems
An Unsatisifed Market
Netflix's recent promotion of their "Open Connect Appliance Software" is just the latest example of the rapidly growing trend of using FreeBSD to build compelling products and services. From the fare meters in New York's taxi cabs, to network and storage appliances used to run the Fortune 500, it's clear that FreeBSD is not just "Turning PCs into Workstations" anymore. Sustaining and accelerating this momemtum requires understanding the changing face of FreeBSD's user base, building relationships with these new consumers of FreeBSD, and working together to make FreeBSD the most compelling and cost effective application platform possible. However, there is a large roadblock in the path to this goal: a credible marketplace offering FreeBSD support and services.
I'll use a hypothetical FreeBSD consumer to help illustrate the problem. Imagine a content delivery business which uses FreeBSD as the OS platform for its servers. Their job is to efficiently distribute lots of data to their customer's end users, via servers deployed all over the world. The in-house staff focuses on developing an application stack to manage the servers, monitor and load balance network bandwidth usage, publish new content, and to reap content that is no longer being distributed.
FreeBSD was chosen by this company because it offers high performance and runs flawlessly on their initial hardware platform. It just works: no need to invest staff time or otherwise suffer distractions from improving their CDN software stack. But initial success does not guarantee long term results. Hardware becomes obsolete. Will FreeBSD fully support the systems available when it's time to replace infrastructure? Deployments grow. Will FreeBSD scale with the business? Operating systems have bugs. How does a FreeBSD consumer protect itself from critical defects?
There are answers to these concerns. FreeBSD has a strong and dedicated community that, more often than not, diagnose and fix issues in hours rather than days. The size of the FreeBSD community means that new hardware support is added to the operating system constantly, often before your business needs these drivers. FreeBSD's scalability has improved in every release and is a focus of the FreeBSD community. For almost any concern raised, an answer of similar form can be honestly given. But these are not the answers businesses are looking for.
Where are the guarantees? The FreeBSD community prides itself on delivering an operating system with mission critical performance and stability, but for many, that doesn't matter without mission critical support. You may retort, "You get what you pay for!" The thing is, these FreeBSD users have money to spend to obtain FreeBSD support. In fact they crave the ability to make it someone's paid job to fix their problems. They just struggle to spend their money effectively.
When a problem crops up, they may try to hire a contractor only to find that the pool of good contractors is small. Not only do they have to find someone in the community who can solve their problem, but there is no guarantee that a contractor is available. And finding one good contractor may not be enough. The perfect contractor for last week's problem may not be an appropriate choice for fixing the problem that cropped up today. A business may incur significant overhead finding, retaining, and managing the contractors they need.
Some businesses hire permanent staff to manage their use of FreeBSD, but this is not without difficulties too. First, hard questions need to be answered. Which areas of the OS will cause difficulty in the future? How many positions are needed and with what expertise? When FreeBSD works well, what should be done with this staff? In reality, the permanent staff model only makes economic sense when the product being developed requires heavy customization of FreeBSD - having FreeBSD experts on staff is already critical to the business plan. Most other FreeBSD consumers would be happy to outsource this function; let an entity that specializes in FreeBSD support aggregate the demand and amortize the headcount costs across several clients. They want a Red Hat or SuSE for FreeBSD.
Regardless of how you feel about the real value of a Linux support contract, the truth is that these services have a large perceived value in the market place. Finance departments and CTOs understand the service contract model. They also believe that a multi-billion dollar company that specializes in support has a better chance of fixing problems than home-grown efforts or community support.
FreeBSD is the passion of our community, but for most of its members, not their job. Having talked with dozens of commercial users of FreeBSD, I've come to this conclusion that this missing economic connection for getting work done in FreeBSD is the largest impediment to FreeBSD's future success. We need to grow the pool of administrators and developers that are fluent in FreeBSD, and we need to build businesses that pay for and deliver these skills to meet the current demand.
Solving the FreeBSD labor problem is a monumental task, but there are steps being taken today to address the issue. The FreeBSD Foundation is working to grow the potential FreeBSD labor pool by bringing FreeBSD into more universities so engineers and adminstrators are exposed to FreeBSD during their college course work. Companies like iXSystems are seeding the FreeBSD services marketplace. The BSD Certification Group provides certification exams for validating expertise on FreeBSD systems. And the FreeBSD community continues to improve its infrastructure so businesses can rapidly find the people and information needed to resolve their problems.
What can you do? Advertise your interest in the FreeBSD platform to vendors, recruiters, and educators to raise awareness of this market need. Invest in the FreeBSD community and FreeBSD education by supporting the FreeBSD Foundation. Patronize FreeBSD support businesses so they can develop their skills and capabilities. If you are a BSD expert, take the BSDA certification exam, or if you are an employer, include BSD certification as a requirement in your job postings. Most importantly, don't wait on the sidelines assuming building a viable FreeBSD marketplace is "not my job." We all have a role to play, and only together can we ensure the continued success of FreeBSD.
Justin T. Gibbs
President and Founder
The FreeBSD Foundation
contributed by Deb Goodkin
I just returned from OSCON, an open source conference in Portland, Oregon. I had the pleasure of working the FreeBSD booth for a day. Since I spend more time running the business side of our company, I don't always have a chance to get out and promote the FreeBSD operating system. I can't even do justice to this after watching Matt Olander enthusiastically promoting FreeBSD. I sat there with my jaw dropped, in total amazement at how Matt, with his true passion for FreeBSD, can open up the most introverted person, and get them talking about FreeBSD or maybe just their experience with open source. When you listen to him explain the history and benefits of FreeBSD, it makes you want to take a FreeBSD DVD and install the OS right then and there. It's people like this, with their incredible passion, who make this such an amazing Project and community to be involved in. And, Matt is just one of the many people I've met, who are improving, promoting, and advocating FreeBSD.
The FreeBSD Foundation's mission is to promote and support the FreeBSD Project and community worldwide. That means we fund project development ($200k allocated), conferences, vendor, and developer summits ($40,000 allocated), FreeBSD infrastructure ($125,000 allocated), travel grants ($20,000 allocated), and other expenses including legal fees, travel expenses for visiting institutions and corporations to help facilitate collaboration with the FreeBSD Project, and producing marketing literature to promote FreeBSD. For 2012, we have budgeted $600,000 to spend on the Project, including our operational expenses, with a goal of raising $500,000.
As of this publication we have raised $180,000. This time last year we had raised around $154,000. We receive most of our funding the last few months of the year, which is very common in the non-profit sector. But, it does make our job of supporting FreeBSD interesting and exciting! How are we going to raise the rest of the funds? Well, first we're going to ask for your help. If you are reading this newsletter, most likely you are passionate about FreeBSD. You can help by making a donation, asking your company to make a donation, and/or spreading the word about FreeBSD and the Foundation. Making a donation is easy. Just go to our donor page at http://www.FreeBSDFoundation.org/donate/ and choose a donation option that is best for you.
Remember, by donating to the Foundation, you are supporting the FreeBSD Project and community worldwide!
contributed by Rafal Jaworowski
It is my pleasure to confirm that in cooperation with The FreeBSD Foundation and Juniper Networks, Semihalf has released to the community the NAND Flash framework for FreeBSD which was developed over the last few years.
The NAND Flash stack consists of a driver framework for NAND controllers and memory chips, a NAND device simulator and a fault tolerant, log-structured file system, accompanied by tools, utilities and documentation.
With the NAND framework available, FreeBSD has been enabled as a platform for embedded and industrial designs where a direct attached NAND Flash is a typical and widespread solution.
The code was initially made public on a SVN projects branch, peer reviewed and discussed over mailing lists, so that we could ensure the NAND framework meets community standards and is easily maintained and enhanced in the future.
It was then merged with FreeBSD HEAD in a series of commits beginning with this one: http://svnweb.freebsd.org/base?view=revision&revision=235537
We'd like to thank the FreeBSD Foundation and Juniper Networks for partnering with Semihalf on this technology transfer initiative. Special thanks go to Semihalf development team (Grzegorz Bernacki, Mateusz Guzik, Lukasz Plachno, Jan Sieka, Lukasz Wojcik) working on this project.
contributed by Edward Napierala
The goal of this project was to make it possible to change the size of mounted filesystems. It's something that's very useful for virtual machines - one can start with a small disk image, and then scale it up as needed. Previously, this required rebooting from another disk. The project involved changes at several levels in the kernel, from disk drivers, to make it possible for the SCSI disk driver to notice that the LUN size has changed, through GEOM, making it possible to resize mounted partitions, to the filesystems themselves. Additionally, a possibility to resize LUNs exported by CAM Target Layer was added.
contributed by Pawel Jakub Dawidek
I'm happy to report that the auditdistd project I was working under sponsorship from the FreeBSD Foundation is complete.
The auditdistd daemon is now part of the OpenBSM package and will be available in its next release.
The auditdistd daemon nicely complements the audit framework. It allows one to distribute audit records collected locally with minimal latency to another system. This helps in postmortem analysis, as we know that at least to some point in time audit logs stored on a separate machine can be trusted. This is very important, because once the system is compromised, we cannot trust any of its local files.
One of the most important goals was to make the daemon very secure. We really don't want any weakness in the auditdistd protocol to allow a break into the machine where audit logs are collected. To achieve this, the daemon makes heavy use of sandboxing mechanisms, including Capsicum, if supported by the operating system.
The daemon can act as a sender, as a receiver, or as both. The whole communication between two auditdistd daemons is secured by TLS encryption. Low latency is achieved by using the kqueue mechanism to monitor local trail files and by sending new audit records as quickly as possible.
For more information on how to setup auditdistd please visit its wiki page.
I'd like to thank the FreeBSD Foundation for sponsoring this project and I hope that it will meet the expectations of the FreeBSD community.
contributed by Bjoern Zeeb
After working on IPv6 for many years now, seeing it finally being used a lot more this year is exciting. Since World IPv6 Launch Day my IPv6-only services are seeing a constant increase in usage. Other companies like Netflix have publicly announced that they are using FreeBSD for their IPv6 enabled "Open Connect" CDN , suddenly giving FreeBSD extra testing pushing lots of bits. In light of all of this the IPv6 performance work has happened at just about the right time.
Since its announcement parts of the work have been committed to FreeBSD and due to a number of requests, the changes will also be part of the upcoming FreeBSD 9.1, allowing users to benefit from it on a supported release. I am very happy that some of the NIC vendors have picked up the hardware assisted offload support for checksums, Large/TCP Segment Offload (LSO/TSO) and Large Receive Offload (LRO) support for IPv6 and provide it with their drivers. Talking to others, I know more are to follow. The maintainers of SCTP have also updated their code to make use of the new features, giving improved performance for that protocol as well.
Having TCP/IPv6 performance on par with IPv4 in the offloading case, allowing for full 10 Gbps line speed connections and UDP performance being closer the to level of IPv4, is a huge step forward.
More patches allowing for better parallelism are in the queue or out for review already, so you can expect further improvements in the future allowing you to also see better performance even without the hardware offloading support. However with the forthcoming release, the focus has to be on a stable stack.
I'd like to thank the FreeBSD Foundation and iXsystems for having sponsored the project, George Neville-Neil for doing reviews, Michael Tuexen for updating the SCTP, Navdeep Parhar at Chelsio for the driver updates, Jack Vogel at Intel, and Andrew Gallatin at Myricom for his help, as well as everyone else who helped or is picking up testing and using IPv6 on FreeBSD.
contributed by Pawel Jakub Dawidek
Capsicum is a novel hybrid capability model that first appeared in FreeBSD 9.0, targeted at application compartmentalization: the mitigation of security vulnerabilities through decomposition of complex and risky applications into isolated components.
Capsicum provides tight sandboxing mechanism where access to all global name spaces is restricted and also allows to limit actions that can be performed on file descriptors (file descriptors represent capabilities).
There are many attempts to implement the former functionality with existing tools, such as chroot(2) and dropping privileges with setuid(2) and friends. Unfortunately this doesn't solve the problem. Such a sandbox can still access various global name spaces, like process table, and can even make internet connections. If an attacker's goal is to build a spam botnet, having machines that can connect to the internet is all he needs. He doesn't need full root access to the system.
The latter functionality is very important as well. Without Capsicum, a file descriptor that is open for reading can only still be subject to fchmod(2), fchown(2) and other system calls.
Capsicum is very powerful, but it requires breaking an application into multiple processes that communicate through some IPC mechanism. This can complicate an application's design greatly.
My goal is to provide an easy to use API that can simplify sandbox management and communication between processes. The easier it will be, the bigger chances to wider adoption and less work to convert existing applications to use Capsicum. At this point I have a pretty complete design of the new libcapsicum API and "Casper" - a system-wide daemon that will provide various services to sandboxed applications, like DNS name resolution for instance.
contributed by Edward Napierala
This was a short project to research what's necessary to develop a kernel-based iSCSI target, based on the CAM Target Layer, and supporting hardware offload implemented by modern network adapters, the iSER ("iSCSI over RDMA," so to say) in particular. It required reading RFCs, experimenting with other implementations to figure out what's nice and what mistakes to avoid, figuring out the organizational mess with iWARP and OFED in general, and glancing at Linux sources. The next step now is to use that knowledge - to actually write the target.
contributed by Dan Langille
For the 7th consecutive year, The FreeBSD Foundation has sponsored BSDCan, an important fixture in the conference calendar. BSDCan started small in 2004 and slowly grew to become one of the leading BSD conferences. It now enjoys a high level of participation from a wide cross section of the global BSD community.
Every year, our biggest costs continue to be travel and accommodation. Without the sponsorship afforded to us by our sponsors, such as the FreeBSD Foundation, we would not be able to retain the talented speakers that we have become known for. Good talks bring more people. More people leads to planning and sharing of great ideas that form the foundations of a great community. Work done remotely is not the same as work done collectively in one location. There is no substitute for good relationships and camaraderie. BSDCan provides a welcoming and friendly atmosphere for this.
contributed by Hiroki Sato
AsiaBSDCon 2012, the 7th BSD conference in Asia, was held on March 22-25, 2012, in Tokyo, Japan. This conference consists of 4 days: a 2-day tutorial/meeting session and a 2-day paper session. There were 8 tutorials (6 in English and 2 in Japanese) and the number of students were 4-20 for each. The paper session had 11 papers and 1 keynote. The number of attendees was 105.
This year's keynote was "Embedded Technology and BSD UNIX in Japan" by Shozo Takeoka, the founder of AXE, Inc., a vendor which provides custom-ordered network equipment based on BSD, middleware for cellphones, embedded BSD, and embedded Linux to major home appliance and digital camera manufacturers in Japan.
In the 2-day meeting session George Neville-Neil and I held a half-day Vendor Summit. About 30 people attended the session and talks were given in English and Japanese by George (past Vendor Summmit), myself (Foundation), Brooks Davis (toolchain), Michal Dubiel (Semihalf), Masaru Oki (Internet Initiative Japan), and Shozo Takeoka (what and why of BSD-BA). Shozo Takeoka, myself, and others are planning to set up an organization "BSD-BA (BSD Business Association)", a Japanese BSD consortium.
Overall, everything went well. And it is notable that there have been several moves to improve the situation of Japanese vendors using BSD. I am also working towards this goal and will keep the Foundation up-to-date.
The tentative date for the next AsiaBSDCon is March 14-17, 2013.
contributed by Gabor Pali
The purpose of the BSD-Days is to gather Central European developers and users of today's open-source BSD systems, popularize their work, and provide an interface for real-life communication. The event therefore features several BSD developers and enthusiasts to tell about their work, communicate with their fellows and reach their potential future partners.
There are no formalities, papers, registration or participation fee, however the invited folks are encouraged to give a brief talk about their favorite BSD-related topic. The language of this event is English, and the goal is to motivate everybody, especially university students, to work with BSD systems. We believe this way developers can get to know each other better and attract more users to learn about new projects and meet the involved parties in person, and eventually join the efforts. Our talks were arranged on a voluntary basis -- motivated by alleviating the amount of administration required for presentation. Instead of having a program committee to set a standard, we are trying to help the speakers in working out their talks if needed.
For this year, the event was co-located with the Austrian Linuxweeks (Linuxwochen Osterreich) in Vienna on May 5, 2012 (Saturday). We had the FreeBSD Foundation as our main sponsor, together with many others who are also closely related to BSD. The support of the Foundation enabled us to cover the incurring costs, therefore to invite 13 BSD committers for a breakfast, lunch, and dinner for the entire day, as an appreciation of their work. We are really grateful for this generous donation.
For the evening before the event, we organized our usual beer meetup (or "stammtisch" as they are called in Austrian) in the downtown of the beautiful Vienna, at a place called Kolar. There we welcomed the incoming visitors and speakers and had a dinner and some beers in a very good mood and discussed the final issues related to the event.
On the next day, the event featured people working with FreeBSD, OpenBSD and NetBSD and touched the following topics: using an embedded NetBSD system to do VOIP calls, introduction to the Capsicum security framework, relayd, the load balancer and proxy solution for OpenBSD (that has been also ported to FreeBSD recently), new features in the FreeBSD ports tree, using DVCSs as cloud solutions, firewalling with pfSense, and finally a presentation on mfsBSD. All of those talks are available on YouTube at the "bsddayeu" channel, and one is free to check out our event photo album available through the web site. We honored each of the speakers with a nice gift package that included a traditional Austrian wine and a Certificate of Appreciation. During one of the afternoon breaks, we even invited our audience to share a cake with us to encourage the feeling of being part of a great and caring family!
As a gratis, Paul Schenkeveld (one of the key people behind the EuroBSDcon Foundation) announced that he created a site for coordinating efforts of European BSD users groups at the end of the event. According to him, it would be a place to collect and maintain event calendars, reports, and pictures of conferences in the region, and potentially share experiences and promotional materials.
We believe that the funds invested from the side of the Foundation will pay off gradually because the more users we can motivate the more donors there will be for the Foundation, and thus for the FreeBSD community. Furthermore, we aim to support the local BSD user groups this way, so they can get in touch with larger audiences in their home country, and make BSD an appealing alternative to local companies. We promoted our event in the BSD Magazine primarily, but our advertisement appeared at many other frequented German and Austrian sites, e.g. heise.de. In addition, because BSD-Day is a multi-project event, we feel that it contributes to a better cooperation between the participating projects, for which relayd turned out to be an excellent example.
We hope to carry on with this successful series next year, somewhere around Central Europe. Do not forget that you can also be part of this enterprise and help us by donating to the FreeBSD Foundation!
contributed by John Baldwin
A FreeBSD Developers Summit was held May 9th, 10th, and 11th in Ottawa, Canada alongside the BSDCan 2012 conference. Seventy FreeBSD developers and forty-one guests attended the summit making it the largest summit to date. Breakout sessions were held on the first two days of the summit on a wide range of topics including Documentation, Offload Engines, Ports and Packages, Virtualization, and the Network Stack. Several developers presented talks during a developers summit track open to all BSDCan attendees on May 11th.
Attendees networked at a casual, summit-wide dinner on Thursday evening, the 10th and stayed up until the wee hours of the morning each day hacking code, discussing ideas, and swapping stories in the Hacking Lounge. The hacking, discussions, and networking continued during the BSDCan conference.
The developers summit would not have been possible without generous support from sponsors including The FreeBSD Foundation and BSDCan. All of the developers and guests who chaired and participated in breakout sessions and gave talks also contributed greatly to the summit's success.
contributed by George Neville-Neil
The Vendor Summit at BSDCan in May of 2012 was our fourth summit to date. The first summit was held in May of 2011, and was followed by a summit in Silicon Valley at the NetApp campus, as well as a summit held at AsiaBSDCon in Tokyo in March of this year.
Each Summit has differed slightly from the previous ones. The Silicon Valley Summit included a full day of Tutorials, taught by John Baldwin and Kirk McMusick. The AsiaBSDCon Summit included presentations by Japanese companies using FreeBSD as well as Semihalf, a company working on support for embedded systems.
This year's BSDCan Vendor Summit followed closely the format set down in the first summit in 2011. Sixty participants signed up representing more than 20 companies as well as the FreeBSD Project itself. For two and a half hours three main areas were discussed: What did the companies have that they could share, what were the things that people were working on which would arrive soon, and what did the project and the companies involved see as necessary to complete in the next six months. A few highlights from the list of technologies or code that were about to go into FreeBSD, and which have now been integrated, include support for ARM 6 and 7, Applied Micro and FreeScale processors, the NAND Flash File System, DTrace support on by default in the GENERIC kernel, integration of the DTrace Toolkit, the ability to grow filesystems at runtime (which was funded by the Foundation), and several more. In all we went through over 100 items on the three lists, and most of these were taking up by individual contributors to make sure they made it into future releases.
The Foundation arranged for dinner to be brought in for all the participants, and the meeting ended just in time, as dinner arrived. Another Silicon Valley Summit is being planned for November of 2012.
contributed by George Neville-Neil
One of the key differentiators for the FreeBSD project is the level of research that is done with FreeBSD itself. All of the major areas of system and language research are well represented in the papers that are accepted for publications by various journals and conferences each year. This is a small sampling of the research carried out on FreeBSD in the last year.
- D. Hayes, M. Rossi, and G. Armitage,Improving DNS performance using Stateless TCP in FreeBSD 9 2010.
- D. A. Hayes and G. Armitage, Revisiting TCP Congestion Control using Delay Gradients, 9-13 May, 2011 IFIP
- L.Stewart, D. Hayes, G. Armitage, M. Welzl and A. Petlund, "Multimedia-unfriendly TCP Congestion Control and Home Gateway Queue Management,"in ACM Multimedia Systems Conference (MMSys 2011), San Jose, California, 23-25 February 2011
- Grenville Armitage, "A rough comparison of NewReno, CUBIC, Vegas and 'CAIA Delay Gradient' TCP (v0.1),"CAIA Technical report 110729A, 29 July 2011
- David Hayes, Lawrence Stewart and Grenville Armitage, "Evaluating the FreeBSD 9.x Modular Congestion Control Framework's Performance Impact," CAIA Technical report 110228A, 28 February 2011 (extended work supported by FreeBSD Foundation)
- M, Tüxen.; Seggelmann, R.; Rescorla,E. Datagram Transport Layer Security (DTLS) for Stream Control Transmission Protocol (SCTP) RFC 6083, January 2011. We used the SCTP kernel implementation of FreeBSD to implement this stuff in OpenSSL. It is now included in OpenSSL 1.0.0.
- Stewart, R.; Tüxen, M.; Poon, K.; Lei, P.; Yasevich, V. Sockets API Extensions for the Stream Control Transmission Protocol (SCTP) RFC 6458, Dezember 2011. FreeBSD was used to implement most of the API.
- Seggelmann, R.; Tüxen, M.; Williams, M. Transport Layer Security (TLS) and Datagram Transport Layer Security (DTLS) Heartbeat Extension RFC 6520, February 2012. FreeBSD, provided the implementation in OpenSSL and tested it on FreeBSD.
- Stewart, R.; Tüxen, M.; Lei, P.Stream Control Transmission Protocol (SCTP) Stream Reconfiguration RFC 6525, February 2012. The implementation experience used in the RFC comes from implementing it in FreeBSD and the API was designed and tested with FreeBSD.
- T.; Rüngeler, I.; Seggelmann, R.; Tüxen, M.; Rathgeb, E.; Stewart, R. Stream Control Transmission Protocol: Past, Current, and Future Standardization Activities Dreibholz, IEEE Communications Magazine, Vol. 49, No. 4, April 2011.
- Adhari, H.; Dreibholz, T.; Becke, M.; Rathgeb, E.P; Tüxen, M. Evaluation of Concurrent Multipath Transfer over Dissimilar Paths Proceedings of the 1st International Workshop on Protocols and Applications with Multi-Homing Support (PAMS), Singapore, March 2011.
- Robert N. M. Watson: New approaches to security extensibility. Technical report UCAM-CL-TR-818, University of Cambridge, Computer Laboratory, April 2012.
- Robert N. M. Watson, Peter G. Neumann Jonathan Woodruff, Jonathan Anderson, Ross Anderson, Nirav Dave, Ben Laurie, Simon W. Moore, Steven J. Murdoch, Philip Paeps, Michael Roe, and Hassen Saidi: CHERI: a research platform deconflating hardware virtualization and protection. Workshop paper, Runtime Environments, Systems, Layering and Virtualized Environments (RESoLVE 2012), March, 2012.
- Robert N. M. Watson, Jonathan Anderson, Ben Laurie, and Kris Kennaway: A taste of Capsicum: practical capabilities for UNIX. In Communications of the ACM 55(3), pp. 97-104, March 2012.
- T. W. Barr, A. L. Cox, and S. Rixner, SpecTLB: a mechanism for speculative address translation Proceedings of the 38th annual international symposium on Computer architecture, pp. 307-318, 2011.
- G. Pali, Declarative Scheduling of Dataflow Networks - Annales Universitatis Scientarium Budapestiensis de Roland Eötvös Nomiatae, Sectio Computatorica, Vol. 37(2012), 2012
Every year we sponsor FreeBSD related conferences and travel to these events for FreeBSD contributors. We believe that BSD-centered and FreeBSD-specific conferences play the dual roles of expanding the FreeBSD user community and supporting collaborative development. The FreeBSD Foundation's travel grant program helps to reduce financial roadblocks to participation in these events.Our grant recipients often send us amazing tales of their experiences, proving the value of this program to the FreeBSD community. You can find these stories and trip reports on our blog.
To find out how to apply for a travel grant, please visit http://www.freebsdfoundation.org/documents/TravelRequestForm.pdf. To get information on how to apply for a grant, please visit http://www.freebsdfoundation.org/documents/GrantRequestForm.pdf.
Here is a list of projects, developers, and conferences we have sponsored for 2012.
2012 Conference Grant Recipients:
- AsiaBSDCon 2012 Conference
- BSDCan 2012 Conference
- BSDDay 2012
2012 Project Grant Recipients:
- Edward Napierala - iSCSI Target Scoping project
- Pawel Jakub Dawidek - Capsicum Component Framework
- Edward Napierala - Growing Filesystmes Online
- Björn Zeeb - IPv6 Performance Analysis
- Pawel Jakub Dawidek - Implementing auditdistd
- Semihalf - NAND Flash Support
2012 Travel Grant Recipients:
- BSDCan - Hiren Panchasara, Mark Linimon, Adrian Chadd, Florian Smeets, Ben Haga, Marius Strobl, Brooks Davis, Julien Laffaye, Warren Block, Daichi Goto, Giovanni Trematerra, Davide Italiano, Thomas Abthorpe
- EuroBSDCon - Brooks Davis
A-Team Systems provides FreeBSD centric server administration and monitoring services across the globe. Our clients know that FreeBSD delivers great performance and rock solid reliability through first-hand experience.
The consistent, unified and stable platform FreeBSD provides allows us to easily deploy and maintain dozens of servers, drastically reducing maintenance time and costs for our clients when compared to other server operating systems.
In addition to server management, we also provide web development and hosting services. FreeBSD gives our low volume clients inexpensive shared hosting without trading off security or performance.
For larger customers, we create clusters using FreeBSD to provide high availability and load balancing. FreeBSD's excellent built in firewall features such as PF, IPFW and CARP make it the ideal server operating system.
When used in conjunction with other open source software, such as Apache HTTPd, FreeBSD virtually eliminates the need for expensive specialized hardware such as load balancers. We've seen great success with clusters using CARP and mod_proxy_loadbalancer on front end firewall and proxy servers with an-easy to scale pool of back-end web and and DB servers.
FreeBSD excels at meeting Internet infrastructure needs, from servers to firewalling. A-Team Systems is proud to put FreeBSD front and center of its service offerings and we are grateful to the entire FreeBSD team for continuing to provide such an excellent platform.
Without FreeBSD, our customers and the Internet at large wouldn't be where they are today and neither would we!
- Adam Strohl, President, A-Team Systems, http://www.ateamsystems.com