A trip to France, or how to be a scientist and a tourist at once

Last month, I’ve been to Nice, France to present an article at the Cloud Computing 2012 conference. It was a rather small conference, with only about 150 people divided to some 7 tracks. The Cloud Computing track was by far the largest, with about 50 people presenting and at least 20 attending every lecture. Although small, it had some markings of a prestigious event – the article acceptance rate was around 30% and the proceedings should be indexed by renowned library services like Thompson Reuters and Elsevier.

Lecture room A

These conferences have several pros about them. Beside getting credit for a scientific publication (in my case PhD student credits, which I need to finish my studies), they are place for meeting new people, ideally interested in something similar as you. I’ve got a list of articles and authors that I found interesting and whom I should contact. And of course, between blocks (containing four presentations (in case everyone was present)), there were coffee breaks, which most people used find and talk to the presenter of the most interesting lecture of that block. So I have already spoken to some of the people from my list right on the spot, and several have sought out me.

A coffee break

A coffee break

Well, and the last good thing about conferences is that you get to see some new place. I imagine that for some of the attendees, Nice could have been quite an exotic spot – there were people from Taiwan, mainland China, Japan and South Korea. But even for us europeans, the French Riviera is a very nice spot to spend a week. Even if for most of the day we were locked in a room with curtains and a dataprojector, I  managed to see all of the city in the evenings and try the sea a few times during the lunch break.



Posted in Theory, Writing | Tagged , , , , , , , , , , | Leave a comment

A small research on.. current cloud research

I was asked to find out what other scientists are doing in the area of cloud computing. So I spent the last week with the CTU library’s Information Resources section. Using the 360Search service I found about 100 articles published in the last 2 years with the keyword “cloud computing”. It’s interesting to point out that the number of articles on this topic published since the beginning of this year is about equal to the number in all previous years. So it seems that this field of research is growing quickly.

The CTU metasearcher can find articles from all the sources that the university has subscribed. The relevant ones were ACM Digital Library with 8 conferences and 14 articles. All of them were indexed by Scopus and some with CPCI-S. Those databases, mainly the second one (published by Thomson Reuters, who also produce the ISI Web of Knowledge), are an indication of the importance of a conference.

Another source is Springer Verlag, where I found 7 journals with 8 papers. As seen from the ratio of those numbers, cloud computing is a side theme in these journals. These journals mostly seemed to be collection of conference proceedings, so they didn’t have an impact factor in ISI, but they had a SNIP number in Scopus (this is a similar bibliometrical indicator of the frequency of citations of articles in a journal, but from a different publisher). Being a marginal source was also true for EBSCOhost with 9 journals and 10 articles. Most of the journals were a kind of metalitarature for librarians, the articles were mostly analyses of application of clouds in publishing and archival.

A much better source, I may even say the best, was Elsevier with 12 journals and 25 articles. All of them had impact factors, mostly above 1.0, and SNIP indexes. I have read 6 articles from one of them and the quality was very good.

Search in IEEE returned 9 conferences with 20 articles. Some of them were found in the citation databases, some not. There was even one conference with the word cloud in the title. IEEE also had three journals with three papers, all the journals highly impacted.

The last big source of infromation relevant to cloud computing is Wiley InterScience with 4 impacted journals and 15 articles. I read 4 articles from there and the quality was also great. After that, there were also several minor sources like MIT Theses, the CTU library and the American Institute of Physics library.

To sum up the research, I present a list of topics that the ~100 articles were about, along with their frequency:

  • 33: Applications
    • 9: science
    • 7: libraries
    • 6: media
    • 4: industry
    • 3: health
    • 2: education
    • 1: enterprise
    • 1: web
  • 18: Security
    • 7: discussion
    • 5: mathematics
    • 4: practical
    • 2: monitoring
  • 17: Infrastructure
    • 11: cloud optimization
    • 4: deployment report
    • 2: architecture
  • 12: General
    • 5: report
    • 4: management
    • 3: green computing
  • 11: Distributed systems
    • 7: cloud HPC
    • 4: parallelization middleware
  • 4: Foundation
    • 2: hypervisors
    • 2: processors
  • 1: Teaching
    • 1: programming
Posted in Theory, Writing | Tagged , , , , , , , , , | Leave a comment

Lecture on cloud computing

Last thursday, I had my first lecture. I’ve given it in the VIA (Internet Application Development) course and it was (you probably guessed right) about Cloud Computing. You’ll probably find the lecture slides on the subject’s website, in case you’re curious.

For inspiration, I have taken the skeleton of a paper I’ve written in the summer. It is titled From Hosting to the Cloud and summarizes the pros and cons of all server, hosting and cloud technologies, from the technical perspective, as well as the economical side.

In the lecture, I first started about servers. The students in VIA are web developers and have so far seen applications run only on their laptops, and the Google App Engine in the better cases. To run your own server, you need to buy the hardware, provide room, power, cooling and connectivity. Licenses for some software are more expensive than the hardware and you need to account for the cost of administration, for even if you are doing it yourself, it is eating your time.

In this part, we had a nice discussion about power consumption and internet lines, which was good, because, as you can see, there is not much hard data in the slides and this way, we discussed some real-life problems.

After that, I started on the theory of Virtualization, its history, the four ways to do it (hardware exceptions+emulation, hardware assisted virtualization, paravirtualization and containers). Then I tried to explain the benefits of Server Consolidation with virtualization. Most of the true web programmers were probably sleeping here, but still, we have talked with some others about CPU types and how to do CPU limiting in Xen.

From virtualization we moved to IaaS (Infrastructure as a Service), which I presented first in its private form, as an upgrade to virtualization, which brings the properties of Cloud Computing to the users server rooms, those are Automation, Elasticity, Accounting and Self-service, and they lead to better utilization of data centers through dynamic re-use of resources and great flexibility in deployment od new and scaling of old applications. I also pointed out that computing power in IaaS comes in units of VM instances, whose computing power cannot be scaled once they are running, so applications need to be ready for horizontal scaling.

I then moved to the public PaaS cloud and it’s Illusion of Infinite Supply, which is great if you need something computationally intensive done quickly without capital investment, but not as good if you need to run a server 24/7, because the operational costs may soon become rather high for bigger loads. But I haven’t written it of altogether. There is a lot of possibilities that can be done with the cloud that are expensive to achieve otherwise, for example high availability with geographic distribution. Or short-term licensing of server software.

I then managed to present the concept of the hybrid cloud and.. I ran out of time. I still had chapters prepared about Webhosting, PaaS and Capacity Planning. Well, maybe next time. Perhaps I will get two lectures next semester, who knows..

Posted in Teaching | Tagged , , , , , , , , , , , , , , , , , , , | Leave a comment

Eucalyptus finally stable

Last week, I finally managed to bring our Eucalyptus installation to a stable state again, after some 2 months of fiddling with it.

Again is the key word here, because we already had it running quite well in the last semester. I have built it in the room E132 as a part time job back then, with some quick node deployment mechanisms thrown in as a bonus. It ran there till the summer. In September we decided to upgrade the capacity by moving to rooms E220 a G3, where new workstations with Xeon processors were recently installed.

As I found out, reconfiguring Eucalyptus for separate CLC (Cloud Controller) and CC (Cluster controller) is not so easy, at least not without complete reinstallation. I’ve had to solve several issues with leftover items in the database, as the master admin didn’t warn me before he unplugged the components. The last of those caused the Storage Controller not to create and attach volumes to instances. I didn’t want to reinstall the CLC completely, because we’d lose user accounts and machine images that way.

And why did it take two months? Normally, one can install a small IaaS cloud in two days time. But that is when he’s working alone and has access to all the components. That was however not the case. The CC now runs on the main server for the classrooms which also runs NIS, NFS and other services for the students. The admin didn’t like the idea of giving me full access to such an important component and thus we had to do most of the debugging by e-mail.

We also had to rewire the network to fully support Eucalyptus’ VLANs, which took two weeks of waiting for some optical transceivers. They say that the cloud spares you the time you have to wait for components when you want to start a new server, but they don’t mention the time you spend waiting for the components of the cloud itself :-).

Now we’re working on getting the Cloud Gunther to a working state again, so that we can resume running computing tasks on it, but more on that later.

Posted in Eucalyptus, Practice | Tagged , , , , , , | Leave a comment

Full Battery Emulator

About two weeks ago, I published some four thesis topics and behold, I already have one bachelor student signed up. His name is Lukáš Vydržel and the topic he chose is the Full Battery Emulator.

It’s not one of my cloud computing topics, but nevertheless, it should be  rather useful. As the reader might know, modern laptops use lithium batteries which contain a circuit called “battery gauge”, which prevents the battery from overheating during charging. Overcharging a bare Li-Ion cell, in contrast to the older nickel and lead technologies, can result in “venting of flame” (R), therefore it can’t be used without a protection circuit. The other function of the circuit is tracking of total and remaining battery capacity.

Seeing how many minutes of power you have left is a good thing, but as soon as you want to do something more adventurous with the laptop’s battery connector, it will pose an obstacle. With older laptops which used a Ni-MH battery pack, powering the laptop off a lead-acid battery was as easy as connecting the positive and negative leads of the battery to the laptop and leaving the thermistor, which was the sole overcharge protection of nickel cells, in place. With the lithium powered laptops, it’s not that easy. The laptop is expecting a power-good signal from the battery gauge circuit and won’t power on until it gets it.

This is where the Full Battery Emulator will step in. In the form of a microcontroller connected to the I2C/SMBus leads of the battery connector, it will implement the Smart Battery Data Specification protocol. It will measure voltage and current at the connector, but it will always report the battery as full.

And what is it good for? Well, there are several applications where you need to power a laptop off a big battery, for example in robotics or wireless networks. There are also people who prefer to carry a big battery around in the backpack so that they can boast about the extreme 10 hour battery uptime. Laptops are usually quite powerful while being energy efficient, which makes them good in all these areas.

The alternative to implementing the Smart Battery protocol is to connect the battery to the power adapter plug on the laptop. To do that, you must use a DC-DC step-up power inverter to get the 20V (the usual voltage of power adapters for laptops) from the 12V at the battery. Those are switching power sources in principle and as such have an efficiency of 50 to 80%.  Yes, and there is another step-down switching power source in the laptop to get back  the 12V needed to power the motherboard. Definitely not a good thing for uptime.

Posted in Teaching, Thesis supervision | Tagged , , , , | Leave a comment