150 likes | 274 Views
ARMD – Next Steps. Next Steps. Why a WG. There is a problem People want to work to solve the problem Scope of problem is defined Work items are defined Schedule is defined Charter on the web http://trac.tools.ietf.org/bof/trac/wiki/WikiStart Scan for ARMD. ARMD Problem statement.
E N D
ARMD – Next Steps Next Steps
Why a WG • There is a problem • People want to work to solve the problem • Scope of problem is defined • Work items are defined • Schedule is defined Charter on the web http://trac.tools.ietf.org/bof/trac/wiki/WikiStart Scan for ARMD
ARMD Problem statement Large Data Center and Multi-Site Data Centers have • Server virtualization introduces massive amount of hosts in a data center • Next Gen or Cloud Data Center have to handle massive amount of subnets (or Closed User Groups) • The topology of subnet changes as virtual machines migrate from one location to another • Traditional VLAN partitions no longer solve problem
ARP issues • 100s Virtual hosts on physical Servers • Sending ARPs • On many VLANs and subnets • Data Centers in single or multiple sites • Multiple sites connected by any type of network technology (L2VPN, L3VPN, 802.1aq Ethernet, TRILL Ethernet) using any type of encapsulation • VM allocation and movement managed centrally or distributed
Problems and Questions Issues • Massive #s of hosts • Massive # of client subnets • VM moving Questions • Can we count on the Gratuitous ARP? In VMs? • How do we minimize flooding of ARP? • What changes need to occur to ARP or Proxy ARP? • Should we focus on v4 or v6 or both? • Can we use the simple solution? • Is there a common ARP for Ethernet (802.1aq, Trill, L2VPN, L3VPN)?
ARP Extensions Extensions to existing Protocols • No requirementany behavior changes on hosts, applications, or Virtual Machines being deployed in the market. • Note: Some solutions suggest • Do not break DHCP, or any other broadcast/multicast mechanism used by applications. • Extensions to ARP and IPv6 ND if possible Retain Common ARP / ND concept • Common protocol across all multiple environments • Adaption in specific protocols (L2VPN, L3VPN, 802.1aq, Trill) Will be family of protocols • At Least ARP
Clarifying Problem • Problem statement • Discussion earlier • Stated in Charter • Questions: • Is this a real problem? • Does the charter statement summarize problem? • Is Security necessary? • Who is willing to help refine problem statements?
Analysis for ARMD (1) • Look for GAPs within Data Center • IPv4 ARP • IPv6 ND or autoconfiguration • Survey existing Work • Survey NHRP (RFC2332) & SCSP (RFC 2334) and application to Ethernet • TRILL work as potential solution • Other existing work from Research (MOOSE), or deployments • Consider proposals • Should consider variety of solutions, including directory based, proxy based, or cache based solutions. • Network or host based
Analysis for ARMD (2) • Analyze Security • Include analysis of security concerns of IPv4 ARP requests from malicious users. Evaluating potential security solutions and conclude if the security threat can justify solutions. • Analysis Security for ND • ARMD assumes the direct links to individual hosts and virtual machines are IEEE802.3 Ethernet links. • Should consider scenarios of one Ethernet network being interconnected by another network, which can be L2VPN, pure IP, Ethernet, or others. • Should consider a performance analysis of proposed solutions.
Question • Is Analysis necessary for these protocols? • Is this the right type of Analysis? • Who is willing to help with analysis documents? • Existing v4 /v6 work? • Security analysis? • New proposal summary?
Protocol • Protocol Documents • Architecture comments • V4 protocol Extensions & protocols • V6 protocol Extensions & protocols • Management • Questions • Do we need architecture document? • Are multiple solutions needed or acceptable? • Is Coordination with Ethernet solutions (L2VPN, 802.1aq, TRILL, Mac-in-Mac) acceptable?
Main 3 Question • Does the community think that given the charter revisions discussed during the BOF (subject to review and finalization on the mailing list), a WG should be formed? • This work is interesting, but refinement of scope is still requierd? • Who thinks a working group should not be formed?