Have Geek Brain

An alternative to distributed caching

There is a technical requirement for easy scaling of the new system. This new system consists of three layered applications (in the form of packet processors). Each level will contain at least 2 servers with the same application on each server.

So, when one of the tiers reaches maximum performance, we can easily expand scalability by adding a new server and the same application to offload some of the processing loads.

The problem is that one or two of the three tiers require heavy caching (about 3 million writes and increments).

I am thinking of using a distributed caching system to solve this problem, but a new distributed caching system would mean an additional error as applications now need to interact with additional caching systems for processing.

I am currently looking at ncache, but just wondering if there are alternatives to this problem? or is there any other comparable distributed caching system that can be similar to or better than ncache and also support enterprise support?

Thanks,

Chen

+2


a source to share


2 answers


You can find in this IBM article (expired) major contributors to DCP (Distributed Caching Platforms) environment.

The alternative we use (not free) is Gigaspace XAP .



http://wiki.gigaspaces.com/wiki/download/attachments/55935974/XAP%20Architecture%20Overview.jpg

+1


a source


Chen -

It looks like you can definitely use a distributed caching system or even in-memory data grid (IMDG). Here are some highlights of Oracle Coherence (formerly Tangosol Coherence):

  • Elastic. Just add nodes. Automatic detection. Automatic load balancing. No data loss. No breaks. Every time you add a node, you get more data capacity and more bandwidth.
  • Use both RAM and flash. Transparent. It's easy to handle 10 or even 100 gigabytes per node coherence (e.g. up to TB or more per physical server).
  • Automatic High Availability (HA). Kill the process, data loss. Kill the server, don't lose your data.
  • Data Center (CA) always available. Kill the data center, don't lose data.
  • The RESTful API is available in any language. Native APIs and client libraries for C / C ++, C #,. NET and Java.
  • Besides simple key cache (K / V), queries (including some SQL), parallel queries, indexes (including custom indexes), rich eventing model (for event driven systems such as exchanges) transactions (including MVCC) are also supported. parallel execution of scalar (EntryProcessor) and aggregate (ParallelAwareAggregator) functions, cache triggers, etc.
  • Easily integrate with your database with read, read, write, and write pass-through. Automatically updates only changed data when changes occur in the database (using Oracle GoldenGate technology).


The Gartner Grid Data-Grid Market Brief is titled Competitive Landscape: In-Memory Data Grids. You can see a copy: http://www.gartner.com/technology/reprints.do?id=1-1HCCIMJ&ct=130718&st=sb

For the sake of full disclosure, I work at Oracle. The opinions and opinions expressed in this post are my own and do not necessarily reflect the opinions or opinions of my employer.

0


a source






More articles:


All Articles