Zigfrid

“Zigfrid” is a platform of creation of high-performance systems for accumulation and processing of data on connection chains between subscribers’ identifiers, formation of historical aggregated databases on relational data with identifiers.

Depending on the type of the processed data on transactions, system configuration and start-up is performed. A series of products solving the most various problems was developed for “CDR” and “IPDR” subject domains.

Purpose:

  • Creation of high-performance real time graph storages for processing of extremely large social graphs and other data on connections of “the identifier – the identifier” type;
  • Creation of real time aggregating databases for the concise long-term accumulation of data, concerning identifiers in transactions, including the periods when accumulation of initial information is economically impossible, and also when formation of the aggregation selection according to initial information takes a long time.

When forming aggregating databases in the automatic mode, only unique data, concerning  identifier information for every day of incoming data, is recorded: users’ ID whom the initial user connected, locations, etc.

Functionality:

  • Search in depth for a given identifier connected on 1st, 2nd, 3d, 4th level;
  • Search of the general ways (connections via identifiers) groups of identifiers via 1, 2, 3, 4, 5 intermediate identifiers;
  • Controlled separation of group identifiers at search algorithms (fraud-numbers, IP-addresses of popular resources, etc.) for exception of insignificant information;
  • Search of graphs without connections with others;
  • Data search by an identifier mask for identifiers in the base;
  • Search by identifying lists and related day aggregated data.

Scheme:

Technical characteristics:

  • Processing of 1 … 10 billion and more transactions a day;
  • Data search in 1 month’s accumulated data:
    – connections on 1st level – less than 1 sec.;
    – connections on 2nd level – no more than 3 sec.;
    – connections on 3d level – no more than 10 sec.
  • Unlimited period of accumulation due to effective compression of the accumulated data (3-8 times), several Tbytes of data for several years of storage are enough for maintaining archive of connections of the large social graph users;
  • Comparable time of search in data for a month, a year, the entire period of accumulation;
  • Standard complete set on the basis of rack-mount 19′ server, 2U assembly size.

Advantages:

  • Minimal expenses for equipment (in most cases one server is enough while popular open source solutions demand clusters and a set of servers for load sharing);
  • Possibility of integration of additional operations, demanding a response in real time, directly into the system core.

Brochure “Zigfrid” (pdf)