Lightweight Internet Traffic Classification - A Subject Based Solution with Word Embeddings

Internet traffic classification is a relevant and mature research field, anyway of growing importance and with still open technical challenges, also due to the pervasive presence of Internet-connected devices into everyday life. We claim the need for innovative traffic classification solutions capable of being lightweight, of adopting a domain-based approach, of not only concentrating on application-level protocol categorization but also classifying Internet traffic by subject. To this purpose, this paper originally proposes a classification solution that leverages domain name information extracted from IPFIX summaries, DNS logs, and DHCP leases, with the possibility to be applied to any kind of traffic. Our proposed solution is based on an extension of Word2vec unsupervised learning techniques running on a specialized Apache Spark cluster. In particular, learning techniques are leveraged to generate word-embeddings from a mixed dataset composed by domain names and natural language corpuses in a lightweight way and with general applicability. The paper also reports lessons learnt from our implementation and deployment experience that demonstrates that our solution can process 5500 IPFIX summaries per second on an Apache Spark cluster with 1 slave instance in Amazon EC2 at a cost of $ 3860 year. Reported experimental results about Precision, Recall, F-Measure, Accuracy, and Cohen's Kappa show the feasibility and effectiveness of the proposal. The experiments prove that words contained in domain names do have a relation with the kind of traffic directed towards them, therefore using specifically trained word embeddings we are able to classify them in customizable categories. We also show that training word embeddings on larger natural language corpuses leads improvements in terms of precision up to 180%.

(Appendix) Grain-size distribution, calcium carbonate content and proportion of cemented aggregates in ODP Site 133-820

The textural and compositional characteristics of the 400 m sequence of Pleistocene wackestones and packstones intersected at Ocean Drilling Program (ODP) Site 820 reflect deposition controlled by fluctuations in sea-level, and by variations in the rate of sediment supply. The development of an effective reefal barrier adjacent to Site 820, between 760 k.y. and 1.01 Ma, resulted in a marked reduction in sediment accumulation rates on the central Great Barrier Reef outermost shelf and upper slope. This marked change corresponds with the transition from sigmoidal prograding seismic geometry in the lower 254 m of the sequence, to aggradational geometry in the top 146 m. The reduction in the rate of sediment accumulation that followed development of the reefal barrier also caused a fundamental change in the way in which fluctuations in sea-level controlled sediment deposition. In the lower, progradational portion of the sequence, sea-level cyclicity is represented by superimposed coarsening-upward cycles. Although moderately calcareous throughout (mostly 35%-75% CaCO3), the depositional system acted in a similar manner to siliciclastic shelf depositional systems. Relative sea-level rises resulted in deposition of more condensed, less calcareous, fine, muddy wackestones at the base of each cycle. Sea-level highstands resulted in increased sedimentation rates and greater influx of coarse bioclastic material. Continued high rates of sedimentation of both coarse bioclastic material and mixed carbonate and terrigenous mud marked falling and low sea-levels. This lower part of the sequence therefore is dominated by coarse packstones, with only thin wackestone intervals representing transgressions. In contrast, sea-level fluctuations following formation of an effective reefal barrier produced a markedly different sedimentary record. The more slowly deposited aggradational sequence is characterized by discrete thin interbeds of relatively coarse packstone within a predominantly fine wackestone sequence. These thin packstone beds resulted from relatively low sedimentation rates during falling and low sea-levels, with much higher rates of muddy sediment accumulation during rising and high sea-levels. The transition from progradational to aggradational sequence geometry therefore corresponds to a transition from a "siliciclastic-type" to a "carbonate-type" depositional system.

A lightweight pneumatic coring device : design and field test /

"September 1982."

Density, absorption and specific gravity tests of aggregates, bituminous mixtures and surfaces.

Mode of access: Internet.

Fine aggregate sources location and geologic characteristics : phase 1 of Physical Research project no. 9 "acceptability tests for fine aggregates" : a report /

Maps in pocket comprise appendix C, Geologic characteristics of fine aggregate sources, and appendix D, Fine aggregate source locations.

Video image analysis of aggregates : a study report on evaluation of coarse aggregate shape and size properties using video imaging /

"Report no. FHWA-IL-UI-278"--Technical report documentation page.

On the seasonal adjustment of economic time series aggregates : a case study of the unemployment rate /

Mode of access: Internet.

Organizational maintenance repair parts and special tools list for electronic equipment configuration for Army helicopter OH-6A (with standard lightweight avionics equipment) (NSN 1520-00-918-1523).

"February 1979."

Operator's manual : tent, general and special use, pole supported, Artic, 10-man, NSN 8340-00-262-2877, assembly, M-1942, NSN 8340-00-262-2877); command post, M-1945, NSN 8340-00-269-1370; hexagonal, lightweight, M-1950, NSN 8340-00-269-1372.8340-935-8372.

"November 1972."

Organizational, direct and general support, and depot maintenance repair parts and special tools list : air conditioner, base mounted, lightweight, air cooled, 38,000 BTU/hr, 416V, AC, 3 phase, 400 cycle, (Harvey W. Hottel, Inc. model HAC-38-416), FSN 4120-788-1163.

"September 1969."

Operator, organizational, direct and general support and depot maintenance manual : air conditioner, base-mounted, self-contained lightweight, air-cooled, 38,000 BTU/hr cooling and 35,000 BTU/hr heating capacity, Fairchild Stratos model VEA4-3A, PN 107201 (FSN 4120-926-4280).

Includes index.