hadoop networking at datasift

18
Hadoop Networking at DataSift "How I learned to stop worrying and love Arista Switches"

Upload: huguk

Post on 13-May-2015

2.438 views

Category:

Technology


1 download

TRANSCRIPT

Page 1: Hadoop Networking at Datasift

Hadoop Networking at DataSift

"How I learned to stop worrying and love Arista Switches"

Page 2: Hadoop Networking at Datasift

About Me

Gareth LlewellynSenior Operations Engineer at DataSift

@NetworkStringabout.me/GarethLlewellynblog.NetworksAreMadeOfString.co.uk

Page 3: Hadoop Networking at Datasift

Overview

● What is DataSift● The DataSift platform in numbers● Our initial network design● Issues with the initial design● Considered designs● Implementation● Questions

Page 4: Hadoop Networking at Datasift

What is DataSift

● Real time and historical curation and filtering of many sources e.g. Facebook, Twitter, YouTube etc

● Augmentation of data e.g. demographic, link resolution

● Real time streams via Web Sockets, HTTP POST / PUT, SFTP etc

● Historical queries against data from as long ago as 2010

Page 5: Hadoop Networking at Datasift

Platform in Numbers; Servers

● ~7k 2.13 - 2.8 Ghz Cores

● ~8Tb RAM

● ~2Pb Storage

● ~380 amps peak draw

● Heterogeneous mix of chassis; Intel SR2600URLXR, Dell R710s & DL380 Gen 7 / Gen8's

Page 6: Hadoop Networking at Datasift

Platform in Numbers; Traffic

● Writes○ ~300Mb/s inbound streams

● Replication○ Peaks of 24Gb/s

● Map Reduce○ Peaks of 70Gb/s

● Exports

Page 7: Hadoop Networking at Datasift

Initial Network Design

Page 8: Hadoop Networking at Datasift

Buffers & Discards

Page 9: Hadoop Networking at Datasift

Moving through the Cisco portfolio

● 2960○ 2.7 mpps○ 32 Gbs

● 3560○ 13.1 mpps○ 32 Gbs

● 3750○ 38.7 mpps○ 32 Gbs

● 4948○ 72 mpps○ 96 Gbs

Page 10: Hadoop Networking at Datasift

Redesigning the Network● Uplink over subscription

○ Servers per cab○ Gbit uplinks per server○ Extensibility / Redundancy of uplinks

● Redundancy of TOR / Core / Distribution○ Power○ Chassis ○ Management Controllers

● Performance○ Buffers○ Head of line blocking

● Extensibility / Scalability○ Number of Hosts / Cabs supported○ Backplane

Page 11: Hadoop Networking at Datasift

Redundancy: Mesh

Pros:● Inter cab transit is truly cab to cab● OSPF reduces admin overhead● Cisco IOS

Cons:● OSPF licence adds cost and

increases complexity● Uplink over subscription● More ports for routing than

hosts

Page 12: Hadoop Networking at Datasift

Uplink Oversubscription: Chassis

Pros:● 720,000,000 pps● 80Gbs of inter blade transit● Cisco IOS● Dual Supervisors / PSUCons:● Still suffers Head of Line blocking● Only 2 PSUs● Overpopulated line cards increase

failure impact● Chassis failure (unlikely) = disaster● And....

Page 13: Hadoop Networking at Datasift

Cables!

Page 15: Hadoop Networking at Datasift

Leaf and SpineArista 7050's & 7048's

● 2x 52x port 10Gbit 7050 core switches● 12x 48x 1Gb / 4x 10Gb port TOR switches ● /27 public subnet per rack● ECMP routes to all racks● Dual PSU with disparate PDU / Dist Board / UPS /

Generator

Page 16: Hadoop Networking at Datasift

The CoreArista 7050

● Each 7050 is a separate layer 2 network● SVI's on VLAN for Internet routing● Static routes● 1.2 Tbs throughput / 960 mpps forwarding

Page 17: Hadoop Networking at Datasift

Top of RackArista 7048

● VLAN number = Cab Number● SVI consumes 1 IP from /27● Static ECMP routes to all other cabs● Minimum of 2x 10Gbit uplinks● 176Gbs throughput● 132 mpps forwarding

Page 18: Hadoop Networking at Datasift

Questions

Yes, we're hiring ;)